Multi Codec Jukebox

Google+ Circles: Humanity's Social Router

2011-07-16T20:02:00.004+02:00

I have been trying to wrap my thoughts around the importance of Google+ circles. The circles idea is to let Google+ users organize their Google+ contacts into different circles. The idea itself is not new; Facebook has let its users organize contacts into different bins for a long time now. The FB avatar of the idea hasn't really been a killer feature; in fact, PC Magazine published a Google+ circles obituary based on the idea's failure in FB.

Most other features of Google+ are powerful and well planned - like video calling (hang-outs), seamless integration with other Google products (Gmail and You-tube), and a good cross-platform HTML mobile app. Still, Google is touting circles as the key Google+ feature. Why does Google think circles is so important?

Lets look at how Google+ circles affects the users social networking experience. By gently forcing the user to select which circle a new contact should belong, Google+ amortizes the job of categorizing contacts. On the other hand new contacts usually end up in one big "friends" bin in FB. The categorization (or binning) has to be performed later (and this is a tedious task - at least I haven't bothered to do it until now).

With Google+, I've ended up with my contacts being in one of these circles:

Fig. 1: Each circle is a post-box to send messages to a specific contact category.

So now I have a bunch of post-boxes, one corresponding to every circle, where I can post information (pictures/status updates/etc.) and they will get routed to that sub-set of contacts which comprise the circle. This gives me the ability to target information to relevant parts of my social network. I look at this as a social graph routing mechanism. Circles are routing rules that users put in place so that their social message streams are routed appropriately in the social graph.

Google+ is constructing humanity's social router via circles which will be programmed via routing rules defined through the elegant circles abstraction. Yes the same thing can be done with FB, but FB never really tried to make this the center-piece of its product. By gently forcing users to separate relationships via circles, Google+ might just manage to make users feel more confident about selectively routing their social lives with different groups of contacts, rather than blasting messages to everyone they (do or do not) know in their huge FB friend lists. The result should be a more information-rich Google+ social network. With more information comes better advertisement targeting possibilities.

Users are more concerned about privacy with respect to their contacts (my family should not see what happened in the office holiday party) rather than Google knowing every intimate detail of their lives. A functional social router will implement this wish without coming in the way of Google obtaining user information. No I don't think Google will be more discerning than FB when it comes to monetizing the private information of users, but hey, who cares about user privacy anyway?

App Engine's Price Shock and the New Web App Equilibrium

2011-05-22T11:57:00.006+02:00

App Engine's Price Shock

The Google App Engine blog announced that Google's App Engine is moving out of beta this year. According to the blog this means that App Engine will have to become financially attractive to Google, and therefore the corresponding upward revision in pricing. The details are intricate and affect customers (app developers) differently based on what services they use; the takeaway is that there is a reduction in the free-tier plan, a fixed monthly subscription fee, and a switch from billing based on CPU cycles to billing based on CPU-instance uptime. Now this has made several app developers unhappy because they feel slighted by Google changing the pricing structure, so much, so late. Google is trying to assuage developer anger, but I am guessing that Google App Engine stands to loose many apps and developers given these higher pricing structures.

Anyway, going back to the 3 highlights of the pricing change, lets see how it improves App Engine's business plan.

Reduction of the free-tier - Google claims that the free tier was eating into App Engine resources because there are many non-paying apps that run just fine using these free resources. Google is out to trim the free tier, leaving just enough to still win over developers who want to experiment with the App Engine, but no more. If an app is seriously deployed, it needs to pay for App Engine real estate. Many apps running on the App Engine are back-ends to iOS and Facebook apps. Why should Google be bank-rolling apps that benefit those platforms? App Engine's early adopter phase is over and freebies can end.
Fixed $9 monthly subscription fee - The App Engine has been a fantastic success with more than 100K deployed apps. Even if 10% of them can be converted into monthly subscriptions then the App Engine will have an evergreen revenue stream (many developers have complained that they would end up pay much more than $9pm given the new pricing model).
CPU-instance time billing - Google wants customers to pay for availability (keeping their web apps running 24 x 7) rather than the amount of computation used by the apps. Since web apps usually remain online throughout, it makes business sense for Google to charge for this 24x7 CPU-instance time. CPU-cycles usage can be little and far-between given the eccentric product life-cycle, usage variability, and popularity of apps. It makes perfect sense for Google to start charging for availability rather than CPU cycles.

Off course these facts are not lost on other cloud providers such as Amazon's AWS service who have always charged per CPU-instance and offer lean free tiers as well.

The Equilibrium Shift

I don't expect developers to shun the App Engine (or other cloud services for that matter) given this and future price increases. Its much more expensive, both in terms of capex and opex, to achieve a cloud provider's level of service availability and convenience for small and mid-sized app shops. But these price changes will affect the technology and architecture decisions of app developers, here is how:

Simplification of server-side logic, and richer client side logic - Clever web app design can off-load CPU cycles to web-browsers, more so with the advent of sophisticated client-side Javascript libraries and HTML5. I expect web app developers to aggressively move more processing into the user browser. This has the added benefit of more responsive web applications, for example HTML5's local data store can be used to store user data and shave-off network latency.
Multi-threaded server-side architecture - Charging for instances would push app developers toward utilizing instances more efficiently, for example, by adopting multi-threaded programming approaches. There is already talk about App Engine developers shunning Python (inherently weak in multi-threaded functionality) and moving to Java.
Space-CPU time trade-off - Storage space continues to be relatively cheap in the cloud. I expect developers to store a lot more application state rather than having to compute it again at a later time. For example, instead of using traditional RDBMS databases (e.g. MySQL), app developers may start looking at simple nosql alternatives like couchDB, which are instead optimized to store multiple indexes and views based on common read patterns.
QoS as a service - App developers will also move toward higher workload thresholds in their load-balancing algorithms, meaning that users may see web apps slowing down as new CPU instances are more sparingly fired up during times of greater demand. I expect app developers to start charging users for differentiated QoS in apps. After all, there is no free lunch, and lunch just got a whole lot more expensive.
Availability as a service - Does it make sense for an app to keep its back-end running (CPU instances online) in the dead of night because 0.1% of its users are insomniacs? I think that some web apps may start experimenting with the breaking the unwritten rule of 24x7 web uptime, or at least charging more for the privilege of using these services outside waking hours. Off course, different time-zones complicate this idea.

In the end, money matters. Its going to be interesting to see the market affecting cloud-deployed web app architecture decisions of the future.

H.264 vs. WebM

2011-01-25T06:39:00.001+01:00

And so it begins. The battle between the H.264 and WebM video codec.

Google's On2 acquisition and the subsequent open-sourcing of the VP8 video codec has created a formidable competitor for H.264. Formidable not because WebM is technically superior to H.264 but because now there is a free alternative to the proprietary and licensed H.264. WebM is free, underwritten by Google, and a proven web-video delivery veteren -after all, Adobe Flash has used On2's codecs for web video delivery over the years.

There are several things going for H.264. First, it is entrenched in several video delivery formats and standards. For example, Bluray uses H.264 to encode video. Millions of Bluray players will become obselete if WebM is used instead of H.264. My two cents are that this wont really happen, instead, newer players will incorporate the possibility of decoding WebM video also. Even as I write this I am aware of several hardware manufacturers who are incorporating the WebM video decoders into their ASIC hardware. But I am not assuming that things like the Bluray standard will be changed, on the contrary, there are other emerging media delivery and storage standards that have been frozen with H.264 being selected as the codec of choice. Standards take years to change or deploy and its very unlikely that they can suddenly adopt WebM instead of H.264.

In the mid-term WebM will defeat H.264 where there is a (easily replacable) software decoder and soft-media. By soft-media I mean video that is not burnt onto read-only media like Bluray disks but instead exists, say, in the form of a web-downloadable video on a server's hard-disk. The economic compulsion of having to pay the H.264 licensing body per-video download and per decoder shipped compared to the free (as in air) WebM alternative shall edge out the former. I suspect web-video delivery platforms like You-tube will lead the charge because (1) The number of videos being downloaded are huge and, (2) Their average revenue per video is miniscule, and each WebM download instead of H.264 download saves a few cents in licensing fee.

A black-knight for the time-frame question will be the innovation in H.264 vs. that in WebM. If open-sourcing WebM has the desired effect of creating a better and more innovative codec in the future then WebM could gain on H.264 faster. But I am sure that the H.264 camp won't be sitting on their palms all this while! Video codecs use advanced algorithms and developing such concepts needs big investments (R&D). Will backers of WebM bring that kind of investment to the table in the interest of improving WebM when there is no direct revenue coming back to them?

Another thing that is going for WebM is the push toward virtualization in consumer electronics (away from the conventional ASIC approach) in the coming years. This means that future hardware (such as future Bluray players) may be capable of running multiple upgradable decoders rather than being tied to a specific ASIC implementing a specific decoding algorithm for a specific codec. That may just break the hardware dominance of H.264 over WebM. As a consumer I would prefer to hedge my bets and buy a virtualization-capable decoder rather than being tied into one video codec via an ASIC decoder.

Parallelizing & Multiprocessing Commands Using Python

2010-11-09T11:51:00.000+01:00

My computer has multiple processor cores. That means I could speed up scripts by running some of their tasks in parallel. I have written up a simple Python script that uses the Multiprocessing library to take a list of jobs (each is a unix command string) and then executes them on a specified number of independent processes. These processes are created only once and act as a pool of "workers" which undertake a job, submit the result of the computation, and then undertake another job (if available in the job queue). The script ends when there are no more jobs in the job queue.
This approach is useful when (1) You have a multi-processor/multicore CPU. (2) Your tasks are CPU intensive. (3) You are reasonably sure that the jobs are not internally parallelized to take advantage of multiple CPUs. In my case, I had two directories full of numerically-named image (.ppm) files whose PSNR's had to be compared using the pnmpsnr utility. Computing PSNR is a computationally intensive task. Running the comparisons serially (single process) was significantly slower than adopting a multiprocess approach.
The code below should get you started on parallelizing your computationally intensive script. You can download the script from here.

#! /usr/bin/env python
# Sachin Agarwal, Google, Twitter: sachinkagarwal, Web: http://sites.google.com/site/sachinkagarwal/ 
# November 2010
# Using Python to execute a bunch of job strings on multiple processors and print out the results of the jobs in the order they were listed in the job list (e.g. serially).
# Partly adapted from http://jeetworks.org/node/81


#These are needed by the multiprocessing scheduler
from multiprocessing import Queue
import multiprocessing
import commands
import sys

#These are specific to my jobs requirement
import os
import re
 
def RunCommand (fullCmd):
    try:
        return commands.getoutput(fullCmd)
    except:
        return "Error executing command %s" %(fullCmd)

        
class Worker(multiprocessing.Process):
 
    def __init__(self,
            work_queue,
            result_queue,
          ):
        # base class initialization
        multiprocessing.Process.__init__(self)
        self.work_queue = work_queue
        self.result_queue = result_queue
        self.kill_received = False
 
    def run(self):
        while (not (self.kill_received)) and (self.work_queue.empty()==False):
            try:
                job = self.work_queue.get_nowait()
            except:
                break

            (jobid,runCmd) = job
            rtnVal = (jobid,RunCommand(runCmd))
            self.result_queue.put(rtnVal)

            
def execute(jobs, num_processes=2):
    # load up work queue
    work_queue = multiprocessing.Queue()
    for job in jobs:
        work_queue.put(job)
 
    # create a queue to pass to workers to store the results
    result_queue = multiprocessing.Queue()
 
    # spawn workers
    worker = []
    for i in range(num_processes):
        worker.append(Worker(work_queue, result_queue))
        worker[i].start()
    
    # collect the results from the queue
    results = []
    while len(results) < len(jobs): #Beware - if a job hangs, then the whole program will hang
        result = result_queue.get()
        results.append(result)
    results.sort() # The tuples in result are sorted according to the first element - the jobid
    return (results) 

 
#MAIN 
if __name__ == "__main__":
    
    import time #Code to measure time
    starttime = time.time() #Code to measure time
    
   
    jobs = [] #List of jobs strings to execute
    jobid = 0#Ordering of results in the results list returned

    #Code to generate my job strings. Generate your own, or load joblist into the jobs[] list from a text file
    lagFactor = 5
    ppmDir1 = sys.argv[1]
    ppmDir2 = sys.argv[2]
    decNumRe = re.compile(u"((\d+)\.(\d+))")
    ctr = 0
    for root, dirs, files in os.walk(ppmDir1):
        numFiles = len(files)
        files.sort()
        fNameLen = len(files[0].split('.')[0])
        for fName in files:
            for offset in range(0,lagFactor):
                fNumber = int(fName.split('.')[0])
                targetFile =  '%0*d' % (fNameLen, max(fNumber-offset,1))
                targetPath = ppmDir2+'/'+targetFile+'.ppm'
                origPath = ppmDir1+'/'+fName
                fullCmd = "pnmpsnr "+origPath+' '+targetPath  #Linux command to execute
                jobs.append((jobid,fullCmd)) # Append to joblist
                jobid = jobid+1

    # run
    numProcesses = 2
    if len(sys.argv) == 3:
        numProcesses = int(sys.argv[1])
    results = execute(jobs,numProcesses) #job list and number of worker processes
    
    #Code to print out results as needed by me. Change this to suit your own need
    # dump results
    ctr = 0
    for r in results:
        (jobid, cmdop) = r  
        if jobid % lagFactor == 0:
            print
            print jobid/lagFactor,
        print '\t',
         
        try:
            print cmdop.split()[10],
        except:
            print "Err",
        ctr = ctr+1
    print

    print "Time taken = %f" %(time.time()-starttime) #Code to measure time

My Grocery Store is a Mobile Operator

2010-10-31T10:40:00.002+01:00

My grocery store sells generic versions of bottled water, soap, breakfast cereal, butter, milk and mobile voice/Internet service. Now thats quite remarkable considering Rewe, the German grocery store chain I am alluding to, doesn't really have a history in the German telecommunications market. What they do have are 15445 stores across Europe that can stock up prepaid SIM cards branded "ja! Mobil" (the name comes from their generic in-store brand). Their physical presence and the mind space ja! occupies drives their business model. If shoppers can drink ja! branded generic cola then they could as well use ja!-branded mobile voice/Internet service.

The innovation here is the marketing possibility offered by Rewe grocery stores (instead of any technical innovation). Rewe has partnered with T-Mobile in Germany to implement its ja! branded "mobile operator". T-Mobile provides a white-label technical platform and Rewe simply brands it "ja! mobile". T-Mobile wins because it gets to sell its service at a discount to lower-paying market segments without putting off the premium T-Mobile customers, Rewe makes a neat profit by leveraging the ja! brand, and the customer wins by getting a discounted service from the best mobile operator of Germany, minus the T-Mobile brand.

I was looking at ja! mobile pricing. There are various flavors of pre-paid and flat-rate plans, although the focus seems to be on pre-paid plans that require no long-term contract and can be dispensed at Rewe's check-out counters. Depending on a customer's typical usage, s/he can can trade-off get a discounted subset of services from among the services offered - SMS, MMS, in-network calling, fixed-line calls, data etc. Interestingly, customer support is not free. Its a little like the contemporary airline business where everything from customer service to carry-on baggage can become a chargeable add-on rather than part of the product. Customers need to be mindful of what their money is buying them before assuming that things like customer service or technical support is part of the product.

Brick-and-mortar stores also sell iTunes gift cards and Facebook credit nowadays. Dell and Amazon partner with Best Buy to sell computers and Kindle e-books respectively. There are interesting business opportunities for anyone who can funnel real customers and subscribers (read: money) into the virtual/communications world. Very real profits await those brick-and-mortar outfits who can build bridges between technology companies and customers, even if they are just plain-Jane grocery stores!

Fancy Vertical Handover: A victim of REST?

2010-10-13T10:00:00.000+02:00

There has been a ton of research, standardization work, and development around Vertical Handover - the ability to change the underlying network access without disturbing the overlying communication protocol (TCP or application) sessions. The simplest example is when a user moves from a Wifi zone (e.g. office) to a 3G zone (outdoors). A seamless handover hides the underlying rewiring of the access and lets the user continue using the device as if nothing changed. Vertical handovers have quickly graduated from laboratory quirk to mainstream occurance, with Wifi-enabled smart-phones switching between access technologies multiple times daily.

But the vertical handover on my smart-phone doesn't really preserve the underlying TCP session and yet works pretty well. Why? Because most of the apps on my phone use REST-ful protocols like HTTP, XML-RPC, or SOAP. That means they are, in theory, stateless. In fact, a TCP connection is created and torn down for every message exchange between the service server and the client. Sometimes TCP connections linger on to improve efficiency (carrying multiple request-response mesages between the client and service server), but a discontinuity in the TCP connection is not catastrophic. I simply see my smart-phone negotiate a new connection with the new access (3G or Wifi) and then my app keeps working as if nothing has changed.

All that talk about preserving TCP connections across access technologies was much ado about nothing!

Mobile Video Calling: Can Tango Tango?

2010-10-06T18:53:00.001+02:00

Tango is a newly launched mobile-to-mobile video calling application for iPhone and Android devices. Tango enables smart-phone owners to see each other in addition to speaking with each other during a Voice Over IP (voip) conversation. Many smart-phones come with front facing cameras, ostensibly for video calling, and Tango enables people to use these cameras during a voip call. Think of mobile video calling when you want to see your expat pet doing silly tricks on video (or for beach and boardroom voyeurism).

But, as Walter Mossberg's Tango's review in WSJ reports, the quality of Tango's video call leaves a lot to be desired. I came across a video on Gizmodo's website showing Tango in action. The verdict is that Tango's performance is way below expectation. In fact, Tango's video frame-rate seemed to be approximately 1 frame per second in the Gizmodo video (and not the "high quality video mobile calling service" as the company's press release claims).

Make no mistake, achieving even 1 frame-per-second video+voice is no small feat. Tango's engineers have packed a real-time video+voice encoder/decoder into a smart-phone and have managed to trasmit/receive two parallel audio/video streams over Wifi (they also claim high quality video calls over 3G but lets not give Tango all the benefit of doubt :-) ). On top of this, achieving this for both the Android and iPhone platforms and for dozens of smart-phone models is admirable.

Frankly, I am not surprised by Tango's dismal video frame rate - resource bottlenecks such as smart-phone hardware, software/OS, network bandwidth and latency have to be overcome before an acceptable double digit frame-rate is achieved. But what surprised me was the poor voice quality: the Tango call sounded a lot like those cheap international calling cards I used to make international calls from the US many years ago. Terrible sound quality. I wonder why Tango engineers didn't trade more video quality (or even cut out video entirely when resources were scarce) and spend resources on improving voice? Voice over IP for mobile phones is a solved problem - Skype and the umpteen number of mobile SIP voip clients got audio to work well even on older smart-phones. Why couldn't Tango?

Tango is an over-the-top application, meaning that it runs over the best-effort (ordinary) Internet. I mention this here because the alternative, 3G telecom-operator-supported video calling, uses a dedicated network channel to ensure call quality assurance. But a Tango call will be carried over the same pipes as plain web traffic, making the video/voice call quality dependent on what else is being transmitted during the call. Telecom-supported 3G video calling is also much more energy (battery) efficient than Tango.Why? Because in order to remain signed-into Tango to receive calls, the smart-phone has to periodically send "I-am-alive" messages to the Tango server. This means that a TCP or UDP socket is always active (or repeatedly created and and torn-down), effectively disabling the smart-phone's built-on power-saving sleep function. Offcourse, telecom supported 3G video calling costs money, but it is technically superior to Tango or any other over-the-top mobile video calling system.

But this is not about Telecom vs. Internet applications. This is about the use-case. Video calling was touted as one of the big use-cases for 3G Telecom networks (and 4G too?). 3G standards support video calling and so there is hardware acceleration, network resource reservation, optimized audio/video codecs, and cross-phone/OS support for video calling on every modern smart-phone. But apart from the cost of making 3G video calls, is their something else that relegated video calling to its sad never-used status in phones? Yes there is. Video calling has simply not been accepted as a viable form of mass communication in our society, and remains to-date, a quirky add-on. When was the last time you placed a video call?

When Internet telephony (voip) arrived it quickly replaced circuit-switched calling. With mobile video calling, even if Tango can eventually fix its technical/engineering limitations, there is nothing to replace! Sadly, the mobile video calling use-case was still-born from the beginning.

Android Device Chatter with the Google Mother Ship

2010-09-01T09:16:00.205+02:00

Parts of this post were moved into a formal research study (click link below).

https://sites.google.com/site/sachinkagarwal/home/publications-talks/gis-2011-infocom-2011

Microsoft Kin: RIP == Social Networks:RIP (?)

2010-07-03T10:34:00.009+02:00

Microsoft Kin 04/2010-06/2010

Microsoft is phasing out its social-network/cloud storage-heavy Kin smart phone just 2 months after launch. This embarrassing report from CNN claims that the Microsoft+Verizon Kin sold less than 10000 units in the two months. RIP Kin.

I never got around to using the Kin, but apparently the market didn't see the justification for the expensive data plan (>=$29 p.m.) tagged on to the Kin by Verizon. The market was supposed to be teens looking to stay connected via social networks, but they did not bite into the insanely high data-plan tariff. Social networking, it seems, is not worth that much to them. How much is it worth anyway?

Lets not belittle the effort Microsoft put into this device - as a product the Kin was fully functional and seemed to do the things you would expect from this sort of device - Internet social networking, cloud storage and syncing of users' data, a built in Zune player, sleek design, etc. And at under $100 (with a data plan) it had a low entry barrier too. It seems like all the pieces were there but the Kin machine never got off the ground.

I don't know if the lack of a credible app store spelt the end for the Kin. What I do know is that social networking apps completely failed to drive sales. Next time someone uses social networking as the use-case for a device or service that is supposed to make money, say - Kin!

Untitled Poem

2010-04-30T18:23:00.004+02:00

Among many other things, my father taught me how to read and write English. Everything I've ever written starts with what he taught me. Now as he lies dying of cancer, I wrote this for him. Say a prayer for him.

All of my thoughts
Like river drops
Together making up me
Like a river that flows
Until it throws
Fresh into the salty sea

All rivers meet that end
No matter what they pretend
Or how many bends they make
And so it will be
With every drop inside me
No matter what path I take

So you may ask
The point of the task
To meander toward the salty end
But don't we all know
Drops become vapor and snow
From which new rivers descend

Plug and play internal HDDs, literally!

2010-01-08T10:05:00.001+01:00

I just saw this contraption on a colleagues desk. As you can see, a 3.5" HDD is literally plugged into the dock as if it were some super-sized memory card . Well, thats exactly what it is. The dock also has ports for USB keys, SD cards, and probably a few other formats.

Interesting to see the form factor difference between the SD slot and the 3.5" HDD slot. Flash memory capacity is quickly catching up with HDD capacity (the latter's lead has shrunk to only about ~10x). HDDs are endangered species!

India's Broadband Future

2009-12-14T10:04:00.015+01:00

Ajit Balakrishnan, CEO, Rediff gave a keynote in IIT Delhi earlier today. His talk suggested that Indian telecommunication operators and the government should not be concentrating at delivering niche multi-Mbps broadband services but should instead concentrate on delivering reasonably good service (100s of kbps) to a larger population. Ajit flashed a slide which showed that 86% of 3G users use their smartphones to access their email, a relatively low bandwidth application, but only 6% use 3G to download and watch videos. Ajit's point was to recognize the importance of broadband as an "always on" connection rather than a high-bandwidth connection in India.

There is a analogue in India's history to this choice that Indian telecommunication operators and the government has to make. The government of India created top notch higher education institutes - IITs, RECs, and IIMs - in the 1950s (after Indian independence). It spends tens of thousands of dollars per year on each student enrolled in these institutes, arguably at the expense of thousands of primary education schools in backward areas of the country. The thinking at the time of creation of these institutes was that this creme de la creme would catalyze the growth of industry and technology in the country. Similarly, it may be theorized that by providing high-speed Internet connectivity, early adoptors will drive applications and create demand in the general population to upgrade their connectivity.

Countries like China or South Korea concentrated on their primary education institutions rather than creating world-class higher education institutes. It is safe to say that both these countries are significantly ahead of India, measured via any human development index. But does this analogy suggest that India should concentrate on democratization of (relatively low speed) broadband rather than creating small pockets of high speed broadband?

I think that the market forces will decide the balance between broadband services in India. The ARPU on low-speed broadband may not exceed $5, but this will be compensated via large volumes. I also believe that low-speed broadband will be served via wireless in India. With mobile phones outpacing fixed line connections by a 12:1 ratio in the country, there is limited scope for technologies like DSL to be widely deployed. Fortunately, 3G, LTE and Wimax are nicely poised to fill in for the lack of fixed line infrastructure in India. As for the niche multi-Mbps broadband, I expect FTTX being deployed in highly urbanized areas where western ARPUs (10s of dollars) are possible.

Thermal imaging cameras at Bangalore airport!

2009-12-06T08:38:00.006+01:00

Photo: Thermal imaging for Swine Flu screening at Bangalore International Airport

Arriving on an international flight at the Bangalore International airport, I was surprised to see two thermal imaging cameras. Each camera was looking at arriving passengers and visually marking those who had an elevated body temperature, in order to discern people who may be suffering from Swine Flu. These cameras are sensitive to IR heat radiation in the body temperature range. The cameras work by mapping temperature readings into a colormap that visually depicts body temperature. The video images produced by the cameras looked eerily similar to the IR images that the alien saw in the Predator movie series!

As compared to conventional body temperature measurements via thermometers, this real-time technique makes it possible for a medical officer to screen many more people. I wonder why these systems are not installed in other world airports.

Multiprocessing vs. Network I/O

2009-11-24T04:38:00.006+01:00

I've been reading up on Python's (v2.6 and above) multiprocessing module. While multiprocessing has been around for a long time, simplified libraries like this multiprocessing module may spur even casual programmers to consider parallelism in their programs. My feeling is that if issues like inter-process communication, synchronization among processes, and deadlock avoidance are dealt with painlessly, then many non-professional programmers would feel confident enough to load up CPUs with programs with multiple processes to speed things up. Moreover, given that multiple CPU cores are becoming the norm rather than the exception on commodity hardware, there is a real incentive to eventually switch to multiprocessing.

What will this switch in program design mean for network data I/O? Will average users end up opening and using more network connections on average? Web browser tabs are a good example of multiple threads or processes. When modern browsers fire up they often connect to several websites saved from the previous session. I conjecture that multiple tabs fill up the network's queue faster than was possible with single core CPUs. Although Network I/O is much slower than CPU bandwidth (data rate at which CPUs process say, HTML), there is a point beyond which a single core CPU becomes the bottle-neck (e.g. firing a dozen browser tabs). But multiple cores remove this limitation and drive network I/O to its physical (or traffic-shaped) limits. I plan to measure this interplay between multiprocessing and network I/O. Watch this space!

Free airport Wifi as a marketing tool

2009-11-12T15:05:00.013+01:00

Google is offering free Wifi in 47 US airports during the holiday season The idea is to flash a few web pages marketing Google's software and services to users in return for free Wifi service. According to this CNN article, Google is not the only company to do so - apparently Lexus and Ebay have also implemented similar ideas, or intend do so in the near future.

Free service is probably going to bring a torrent of airport Wifi users online - probably many more than the current number of (paying) users. Given that Wifi Internet channel space is a shared resource, it will be interesting to see how airport Wifi scales with the up-tick in usage. I just hope that the service doesn't deteriorate so much that the sponsoring companys' well-meaning message is lost to disgruntled users. And I do hope that engineers running these Wifi access points have done the networking provisioning Math beforehand.

Now the economics. The sponsoring company (Google) is probably going to pay a lot less than the retail price of airport Wifi connectivity. Why? Because the sheer volume of users will be much higher than when users have to pay individually. I think that the payment will include a fixed component depending on the number of access points participating in the service, and a variable component depending on the number of users accessing the service.

Lets assume that an average airport has about 20 accessible Wifi access points. Each access point can support (with any reasonable quality of service) about 10 concurrent users. If the airport is busy for, say, 12 hours in a day, and further say that we assume an average utilization of 50% of the total capacity of the access points, then we have (per day)

10 * 20 * 12 * 0.5 = 1200 hours of usage per day per airport.

I would assume that the sponsoring company (Google) would pay about $5000 per day as a fixed cost and then about $1 per hour usage. This brings the daily total cost per airport for the sponsoring company to $5000+$1200 = $6200.

So for 47 airports and 50 holiday season days, we are looking at a bill of about

6200 * 47 * 50 = $14.57m

That's not a bad deal for a big company like Google, considering the number of eyeballs they will capture. Lets say a user uses the free Wifi for 30 minutes on average. So, we are looking at about 12*10*20/(1/2) = 4800 users per airport, per day. That works out to over 11m users in the 47 airports over the 50 day holiday period. Even if we assume that most people make round trips and therefore use the Wifi connection 2 times, Google can still reach about 5.5 million unique users! No too bad for the $15 million spent.

And I haven't even started counting the goodwill ROI bonus for playing Santa during holiday season! Nifty nifty marketing.

Call for action! Powering down PCs

2009-11-06T13:51:00.010+01:00

I've been playing with the idea of building a PC application that measures a computer's idle time. The idea is to gently convince users to suspend or power-down their PCs when they are not being utilised. I strongly believe that if PCs are optimally powered down, then many users could cut energy consumption (and hence also save on energy bills). Powering down battery-powered laptops will also increase the longevity of batteries and thereby decrease toxic battery waste in landfills.

As an example of where the possible savings may be, above is a pie chart showing my own PC usage over the past few working days. As you can see, there is ample scope to power down/suspend PCs when they are idle.

If you want to contribute time to this project (coding/web page/translation into other languages/spreading the word), feel free to contact me. If not, then do suspend your PC every time you are away for more than a few minutes :-).

Impact of International Domain Names

2009-10-30T23:03:00.012+01:00

On the 40th birthday of the Internet last week, The Internet Corporation for Assigned Names and Numbers (ICANN) formally announced that there would now be domain name support for non-latin character URLs. This concept, called international domain names or IDNs, will allow URLs composed from letters of scripts of languages such as Korean, Chinese, Hebrew, Arabic and Hindi.

A little digging on Wikipedia about IDNs reveals that the underlying implementation is based on translating unicode names into DNS-compatible (ascii) URLs and visa versa in order to keep the current DNS system functional. This makes the system backward compatible with currently deployed name resolution infrastructure. In fact most of the translation to/from the non-Latin scripts will be done on the users' browsers.

But what does this mean for the fabric of DNS address space and the web?

Dilution of Latin namespace (?) Will we see some dilution in the value of address real estate? For example will http://www.doctor.com become less valuable because folks in Germany can now remember it instead as the more meaningful http://ärzt.com (Ärzt is German for medical doctor)? And what of those tens of thousands of domain names registered from different languages in Latin script (e.g. http://naukri.com in India. Naukri in Hindi means job).

The registration rush Initially, web content providers will scurry to buy up non-Latin names. But this will be more important for those content providers who do not have a global brand-name, or have a brand-name that defines their product or service. For a content provider like doctor.com, it will make sense to buy the synonyms of "doctor" in other languages, in addition to the spelling of "doctor" in the other languages. On the other hand, Microsoft.com will only buy up the spelling "Microsoft" in the languages/scripts becoming available through IDNs. At the very least, I forsee most businesses re-evaluating their namespace position on the web.

Security and phishing Completely unrelated characters in different scripts can look the same to the human eye. This means that users can be tricked into thinking that the address displayed in the address bar points to a legitimate page when in fact it points to a phishing page. It may be prudent for businesses to be aware of these security vulnerabilities of their URLs and perhaps register "similar looking" URLs in other languages/scripts proactively.

Impact on search engines Search engines are known to weigh in address name strings in their ranking algorithms. This may need some re-thinking. At the very least, some search engines may need to use automated translators to link up semantically similar web pages irrespective of how the address space links different copies of the same information in different languages/scripts.

Feel good sustainable energy video from RWE

2009-07-25T20:08:00.004+02:00

This video from RWE, Germany's second largest electricity producer, is sweet. Too bad that in reality, Germany depends on polluting coal for much of its electricity.

Pimped up Tom-toms...Netbooks+GPS: where is the value?

2009-07-05T22:53:00.010+02:00

There is something funny going on in consumer electronics and personal computing nowadays. There is the tendency to combine different functionality into one all encompassing device. Management calls it convergence. Sales and marketing call it up-selling. The product guys call it how-to-keep-your-job-important. Consumers see it as a peculiar phenomena where devices become pricier even though they should be getting cheaper.

Here is a good example :

Smart phone = phone + camera + GPS + DVB-H TV receiver + memory stick + music player + ...

OK so I see some value in this smart phone convergence algebra. All the above services may be useful to a user and its great to carry them all in your pocket.

But it struck me as odd that Dell is planning to up-sell its Mini notebooks by charging users for extra GPS hardware. Apart from robotic aficionados who would want such GPS service on a netbook? The significant time it will take to power on a netbook and fire up the GPS would probably mean you've overshot a couple of highway exits before the GPS locks on and tells you where you are. And that will be after you've somehow placed the netbook over the dashboard for line-of-sight to GPS satellites, driving at 70mph. Netbooks with their limited battery lives would make for poor trekking-in-the-Rockies aids (and why would someone carry a netbook instead of a small portable tom tom up the trail anyway)?.

But the guys at Dell must have thought of all this. Even after you remove the management, sales, and product-team views on such matters, there must be someone who tried to write up credible use-cases. What is the killer-app use-case?

In my opinion, there is a very significant business-opportunity here. Dell wants to know where you are, in order to introduce location-based services (e.g. locality-aware advertising) in return for GPS map services. If you think about it, operators have an advantage over other IT vendors because they know (through cellular triangulation) where a user is. Hardware and software vendors would also want a piece of this action because of the significant scope for location-based advertising. GPS gives them that chance (and more accurate location information). Dell could know, through its GPS/map services, that its 13:00 on a Wednesday afternoon and I am sitting in a downtown Berlin park about 3 minutes walking distance from a restaurant serving schweine haxe and Berliner pilsner. I bet I'd take the bait of a 10% coupon to get there! And Dell would get a piece of the pork too.

P2P, bandwidth, and FTTH urgency.

2009-07-02T11:31:00.059+02:00

Figure 1 (from MPI-SWS) Bittorrent throttling by geographically spread ISPs. Red areas indicate ISPs throttling Bittorrent traffic.

This figure is from the Max Planck Institute for Software Systems' Glasnost project. It shows geographical regions where ISPs interfere with Bittorrent traffic. Comcast (and several other ISPs) claim that P2P applications of a few users slow down the Internet for all network users. All the bandwidth is used up by a miniscule subset of the subscribers, leaving everyone else with a slow network. Theres no reason to disbelieve this argument, a limited shared resource being over-used will result in poor quality for all users in the statistically multiplexed Internet.

There are 2 ways of dealing with the issue. Either bandwidth-hogging users are cut-off (like Bittorrent throttling), or, the network capacity is increased to accomadate the "over-use". The latter technique bailed us out the last time Internet traffic exploded. Broadband was roled out just as media rich Internet applications were catching on (or was it the other way?). Everyone was happy. Customers got better service for similar subscription costs, Web 2.0 companies got the pipes to deliver their content, and ISPs created the whole broadband business, with the option of up-selling through services like digital IPTV.

If it worked so well at that time, can't we do the same trick again? Why not roll out broader-broad-band: FTTH (Fiber to the home) for example? The simple answer is that we cannot, at least not quickly, given the cost. When broadband came the physical access network was already built! There were cable TV wires running to homes and there were phone lines. In terms of a tree analogy, the leaves of the access network were already connected up. All that remained to be done was to put in the trunk links and the branches. And there are a lot fewer branches than there are leaves. On the other hand, FTTH will be prohibitively expensive in many countries - the leaves need to be rewired. Therefore the rollout timeline is going to be slower as compared to broadband.

Back to P2P. Why single out P2P? Don't video CDNs like You-tube, Netflix, Hula etc. also consume large amounts of bandwidth? In my opinion the extra load on the access network imposed by P2P, due to the uploading aspect, creates a lot more congestion in the access network at present. A P2P system will upload (in theory) as much as is downloaded in the system. And all this happens on the access network. Thats a 2X increase in bytes traversing the most expensive component of the network (the edge). This means many many more expensive boxes to cover the leaves of the ISP's tree.

However, most broadband connections are asymmetric (downlinks have higher data rates than uplinks). So P2P is limited to a glass ceiling (the lower uplink bandwidth data rates). On the other hand, conventional CDNs push data down the wire, and so, there is no reason for CDNs to limit video quality and resolution until they hit the downlink rate. As high-quality video content catches on, there will be disgruntled users who wonder about the difference between what data-rate their subscription plan claims (XX Mbps) and what comes down the wire (XX/ZZ Mbps, ZZ being the over-subscription factor).

ISPs need to hurry up fiber-wiring up those leaves! And governments need to help with the capex! Another stimulus perhaps?

Twitter's Value Proposition

2009-06-21T21:52:00.055+02:00

Figure 1: Top-20 query set similarity, or, what fraction of the top-20 queries, separated by a time lag (X-axis) are the same. (Click figure to enlarge).

Figure 2: Frequency of occurrence of a query name vs. the number of unique query names in the top-20 (hourly) query sets over 6 months. About half the queries appear twice or once (Click figure to enlarge)

Twitter has become the latest darling Silicon Valley start-up. What started off as a seemingly incremental idea ("micro-blogging" in 140 characters or less) seems to have caught on big time - more people are twittering than ever before. Twitter has shown its prowess in everything from influencing the American presidential election to challenging Iranian theocracy. Its popularity makes it a very compelling service, but how can it make money for its founders, Evan Williams and Biz Stone, and its promoters?

Needless to say, many Twitter posts (tweets) are inane (A dog bit me) instead of newsworthy (I bit a dog). Many users underestimate the difficulty in producing a constant stream of interesting 140-character long information texts from their everyday lives and experiences. Fortunately, Twitter comes with a search engine optimized to index tweets in real-time. So users can query Twitter for useful information within the deluge of tweets being posted every moment. This makes Twitter a real-time application and a perfect vehicle for propagating news on the Internet. Product releases, reviews, security bugs and vulnerabilities, company press releases, executives' and analysts' statements, etc. previously had to wait for a search engine to crawl and index them on the Internet (a lag of weeks sometimes). But because the tweets first come to Twitter directly from users, these announcements are instantly indexed and available for search via Twitter's search engine.

Twitter provides a great API to study queries returned by Twitter over a time period. The API allowed me to download information about the top-20 most-popular queries submitted to Twitter in every 1-hour interval over 6 months. Parsing this information sheds light on what users search for when they go online looking for time-sensitive information. It also suggests ways of monetizing this vast treasure of users' mind-space - what they are think, search, and find as time goes by.

Figure 1 shows top-20 query name set similarity, or, what fraction of the top-20 queries, separated by a time lag (X-axis) are the same. Twitter reported the set of 20 most-popular query names in each hour. The figure is plotted by finding the fraction of common elements between any two such sets separated by a certain lag (X-axis). The similarity quickly dies within 48 hours, and after 14 days, it settles at about 0.1 (meaning only 2 of 20 query names remain the same between the compared query name sets). There is also a noticeable drop in similarity at 24 and 48 hours (probably due to periodicity effects). Also note the diurnal bumps. Why does that happen?

Figure 2 plots the frequency of occurrence of a query name vs. the number of unique query names in the top-20 (hourly) query sets over 3 months. The majority of top-20 queries captured users' interest for a very few hours. Only about 20% of the queries remained in the top-20 lists for more than 10 hours in the 6 month period.

But what does all this mean for making money via Twitter. Well, here are my 2 cents:

1. Marketing benchmarks: Companies can use Twitter query-popularity to measure their marketing success. Say, how much has an advertising campaign been able to enter a customers mind measured in terms of query frequency (say as compared to their competitors). Twitter can develop and sell analytic tools for companies to measure such stuff. In my opinion, looking for this information in user queries is more effective than looking in the tweets themselves because the latter can be gamed (spam tweets) and because tweeting users are still a minority of people passionate about posting messages (as compared to the silent majority that does not post). Monitoring product mentions in query names is also a great way to keep tabs on marketing success. For example, in Figures 2 and 3, a product name occurring frequently is good news, but if the rate of occurrence decreases over time then its time to launch more marketing efforts or to improve the product's visibility in some way.

2. Real-time Customer feedback: Product groups can use Twitter query information to pinpoint product bugs, fast. There is a certain cost for a user to go on the Internet and search Twitter for say, "iPhone screen blank". If such a query bubbles up in query popularity (say, a top-XX query), then the bug is most certainly a widespread issue. Twitter's real-time feature highlights problems very quickly and efficiently. Selling such product specific query information to companies may create a nice revenue stream for Twitter.

3. Keyword Analytics: The occurrence of a product name with another query word may signal a selling opportunity. For example, the query "iPhone anti-virus" may point to market demand for Apple to sell anti-virus software with its iPhone. Keyword analytics can also be used to help in choosing keywords for online advertising.

4. Risk Management: Twitter quickly captures the viral spread of information on the Internet. This could allow a company to react to, say, a malicious video posted about its product on You-tube. Twitter is your fast-response Internet guardian. For example, Twitter can offer a service to subscribed companies that notifies them about any information (positive or negative) gaining traction on Twitter.

And finally, here is a list of Twitter search strings that were in the 20-most popular search lists for 100 hours or more (over a 6 month period). Funny how Apple dominates the top-3 slots, and then there is AT&T on the 4th slot (probably due to selling the iPhone in the USA). Hats off to Apple's marketing to have captured so much of users' mind-space. Or are they gaming Twitter? Or are Twitter users disproportionately Apple fans? Or is Twitter the newest Apple rumor spreading mechanism?

iphone	3266
ipod	1715
apple	1711
at&t	1253
itunes	981
goodnight	975
tweetdeck	881
vegas	680
bbc	674
new york	645
texas	643
obama	529
star trek	503
gaza	491
lakers	485
susan boyle	484
swine flu	460
sxsw	452
slumdog millionaire	448
watchmen	437
iranelection	427
h1n1	398
dollhouse	394
american idol	382
tgif	364
musicmonday	359
easter	354
wolverine	331
kobe	330
adam lambert	320
windows 7	320
valentine's day	281
aig	275
tehran	271
swineflu	267
bsg	259
sydney	256
super bowl	251
starbucks	242
palm pre	224
twilight	223
christmas	223
austin	217
mexico	216
coraline	212
california	212
cavs	206
nba	205
oprah	204
ces	198
ncaa	197
follow friday	195
spring	194
iran	192
wii	190
transformers 2	188
	188
miami	186
bush	183
paris	179
mousavi	177
snl	172
social media	171
spotify	170
superbowl	164
jimmy fallon	164
diddy	164
imax	160
hamas	156
inauguration	154
heroes	154
twestival	153
blackberry	149
brazil	148
macworld	147
earth hour	145
florida	144
lebron	143
happy new year	143
chuck	142
safari 4	138
snow	138
mj's	137
benjamin button	136
bing	136
chelsea	133
chris brown	133
gran torino	130
ellen	127
kris allen	126
nascar	125
teaparty	125
grey's anatomy	124
wimbledon	124
president obama	123
canucks	122
uksnow	122
ted	122
fridays	121
therescue	119
michael	118
michael jackson	118
conan	117
oscars	116
liverpool	116
celtics	115
angels & demons	113
march madness	111
denver	111
blackout	110
jay-z	109
google wave	109
dallas	107
memorial day	106
xbox live	104
gmail	102
player snapshot	102
north korea	102
mardi gras	102
melbourne	101
french	101
father's day	100

Sun setting over the German countryside

2009-06-01T22:25:00.001+02:00

Took this one while sitting in the ICE from Cologne to Berlin, between Wolfsburg and Berlin Spandau railway stations. 01.01.2009, 21:01 CET.

The hope of green power ?

2009-06-01T22:25:00.000+02:00

Is that Sun driving the wind driving the power-generating windmills producing electricity to power this ICE train cruising at 180 kmph? Please let this be the future for all our sakes.

Is P2P dead?

2009-05-03T13:48:00.011+02:00

There have been some significant setbacks for P2P in the past year or so.

Pirate Bay's founders are in jail in Sweden for abetting illegal file sharing on their website.
Joost, the much lauded P2PTV service, is no longer P2P but is instead a CDN-type streaming service.
Another P2P darling, Skype, seems to be adrift with Ebay wanting to get rid of it through a sale or a spinoff.
There is a sustained reduction in CDN costs that is making the P2P cost reduction less attractive.
Websites like You-tube won the video streaming battle against P2P video streaming long ago, now websites like Rapidshare are leaving P2P file sharing behind as well.
Mobile P2P (near-network P2P on mobile devices over Bluetooth etc.) just didn't happen. These devices are more client-serverish than wired devices because upload bandwidth needed for P2P is too pricey.

So the key question is whether P2P is dead.

In my opinion, the answer is No. Here are some reasons

P2P lacks a business model but has proven to be a remarkably resilient and cost-effective technology. The problem is getting legal content on to P2P networks. Content companies are not going to let users take control of content delivery.
But if one is to look at where the biggest growth in broadband usage is going to be, one looks toward China and India. The legal protection for content is significantly weaker in these countries. Moreover, there is a large amount of reasonably priced content (e.g. regional and Bollywood content in India) that will perfectly ride P2P networks.
P2P has proven itself for voip (Skype has 400m users). Skype is the established voip leader and it will remain that way for a long time.
CDNs do not scale with video quality. That is why Youtube won't do HD - they'll go broke paying for CDN (server) bandwidth. P2P on the other hand can scale up to the extent the access networks allow.

What we may see is an amalgamation of CDNs and P2P technologies. For example, using CDNs for paid content and P2P for promoting the paid content (for free). Cache every recent movie's first 10 minutes on a user's computer using P2P, and then stream the content the user selects via a CDN. Although content pricing can recover CDN distribution costs, the monetary transaction only happens after a user selects to watch. Prior marketing can be P2P.

Pinchpoints in non-linear growth businesses

2009-04-22T18:47:00.016+02:00

Non-linear growth business is best explained via examples. Twitter is a non-linear growth business because the number of users doubles every few months. Cellular service is a non-linear growth business because the number of cellular phone subscribers worldwide has doubled every year. Computer hardware is a non-linear growth business because the number of computers has doubled every few years and Moore's law has doubled the number of gates in a microprocessor every 2 years.

Naturally, such non-linear growth in a business attracts competition. For example, there are plenty of social networks vying for user attention and competing with Twitter. There is healthy competition in cellular service providers in most countries. Similarly, IBM has been supplanted in the PC hardware business by HP, Lenovo, Dell, etc. So the profit in a non-linear business can quickly evaporate due to competition. The pie is divided up. Capitalism and market forces 101.

But there are companies that build effective moats around their products and keep the whole pie. The moat is usually a combination of two components. One component is a standards and IPR protection for products. By creating a standardized and widely adopted product (e.g. MS Windows) companies can effectively keep others out. The second component is the high entry bar. A modern CISCO router ruitinely runs a 15-20m line operating system, making it incredibly hard for a start-up to join the fray.

CISCO operates on a pinchpoint, at the bottom of an inverted pyramid. At the top are all Web 2.0, Internet applications, and data centers. All the growth and innovation at the top makes money for CISCO. Intel is another pinchpoint and all PC manufactures compete at the top of the inverted pyramid. Truly smart are the business models that perch companies on the pinchpoints... the pie just keeps getting bigger!

RIMM and blackberries doing just fine

2009-04-13T20:12:00.009+02:00

I remember thinking a couple of years ago that Apple iPhones were going to knock the Blackberry off the smart-phone market. But nothing like that has happened. RIMM released some very convincing quarterly results last week. Apparently RIMM seems to be doing just fine, never mind the slew of smart-phone entrants. On top of protecting its corporate user turf RIMM is fighting back with the Blackberry Pearl and Curve, taking the battle to the consumer smart-phone segment. Look out Apple (and Nokia), here come the smart-phone mounties!

RIMM has taken the Apple device threat very seriously. Their device line-up has been continuously updated and has kept pace with the latest cellular technology and device features - 3G, application-store, GPS, cameras - you name it and there is a Blackberry model covering the area. Even user interface, Blackberry's Achilles heel, seems to be getting a lot of attention on the newer models. Its a good thing RIMM is looking beyond push email and into actually building general purpose smart-phones. Smart-phones are good for the company's bottom line because the lions share of RIMM revenue comes from device sales. Users need to see glitzy, application laden devices to consider upgrading frequently. After all email by itself is only a text-based application (remember those monochrome Blackberries that handled email just fine). Fortunately, RIMM has not fallen into the Polaroid-like trap of living off one single success and has instead covered broad market segments with multiple device offerings and continious innovation.

Fortunately for RIMM, Apple has built a deep moat around its iPhone through aggressive pricing and exclusive deals with Telcos that keeps many corporate buyers in Blackberry's stable. Example: Small corporation X, 2000 employees, per unit iPhone cost minus per unit Blackberry cost = $250. Choose the Blackberry and you've saved half a million upfront, avoided iTunes (consumer software) being downloaded and installed on all corporate PCs, and escaped the clutches of very expensive iPhone data plans.

Blackberries are great texting/messaging devices with physical keyboards unlike iPhone's glass tapping virtual keyboard. This qualifies blackberries as the serious corporate tool as compared to the iPhone that has yet to fit into the corporate IT comfort zone. And this brings me to the greatest strength of RIMM - integration with corporate IT systems.

The key Blackberry value is effortless outsourcing of wireless/mobile email for corporate IT departments. No extra servers to run and no expensive data-plans with Telcos. Just a per-month per-account fee and you've got wireless email for employees. Since every other corporation (and Barrack Obama) trusts the Blackberry service, so can corporation X. Outsource email with confidence.

One of the greatest shortcomings of corporate PC email has been the culture of creating and maintaining individual email systems per corporation. This system could have been much cheaper, much better managed and less buggy had there been a uber-email provider like RIMM is for corporate wireless email. And no I don't buy the argument that corporate email needs to be kept inside the company's intranet for data security reasons. If this is so important then please forbid employee blackberries which otherwise bounce every email off a server in Canada.

Fortunately wireless email has remained concentrated in the hands of Blackberry due to its early dominance in this technology and this has made RIMM the single largest wireless push email provider. Everyone benefits, even competing smart-phone manufacturers who can simply buy the Blackberry service for their devices. And thats why RIMM is a buy.

Why bonuses should not be curtailed in banks

2009-03-20T15:10:00.014+01:00

There is a flurry of law-making activity in the US House of Representatives and in the US senate to impose upto 90% taxes on large bonus payouts to individuals. AIG's ill-timed bonuses have outraged the entire nation and have led lawmakers to to use this blunt blow to root out bonuses in all sick companies supported by taxpayer money. But the move is a populist one, and one that may have far deeper consequences than recovering a few hundred million in bonus money. I am afraid there is a lot to lose with this last minute decision of making a law limiting bonuses. The taxpayer will actually end up loosing much more in the long run.

In many companies bonuses are given across the organization to all employees. A bonus for a good quarter, a bonus for Christmas, etc. But in knowledge-intensive fields bonuses are disproportionately distributed and given a small subset of employees who create exceptional (significantly over the average) value for the company. Knowledge-intensive fields include finance and banking where a talented person can create a lot more value than another lesser talented person. The distinction between employees is easily measurable in banks (earned profits) and therefore banks can easily reward better-performing employees. The reason for the reward is the hope is that a better performing employee can be retained in the company if she is paid a bonus.

A modern bank's success is almost wholly dependent on its employees. Banks are no longer armies of clerks following a rule book and doing the same thing every single day. Instead, banks are participating in a very sophisticated knowledge-based global game where bankers make decisions on how to invest their depositor's money. They employ some of the brightest minds to compete with other bright minds of other financial institutions, all of them trying to maximize returns on invested money. Quite naturally, the financial institution with the most creative, innovative, smart, and industrious people will take it all away.

Not all minds are equal. Hence all bankers cannot be paid the same income. Its a mind game, and intellect is not equal across the board in every employee (perhaps unfortunately, but certainly by natural design). Bankers are men and women like the rest of us. They will gravitate to financial institutions which pay better. By removing the bonus incentive for bright people in ailing financial institutions we are clearly reducing the chances of these institutions recovering taxpayer money as the brightest will leave or be less motivated.

Bonuses have been wrongly maligned in this whole saga. They are one of the best tools to improve worker productivity, commitment, and a company's bottom line.

East Berlin

2009-03-12T20:58:00.000+01:00

And a not-so sunny day, this time the view of East Berlin.

Firefox: Finally, a credible threat to Google?

2009-03-10T21:36:00.008+01:00

Today there is an extremely interesting article in WSJ about Firefox's threat to Google. No this is not about Firefox vs. Google's Chrome browser. It is about a small feature in recent Firefox browsers that stores web-pages accessed via the browser off-line and presents this locally-stored content as search results to users as they type words into the Firefox browser's address bar. Such local caching is not a new concept, but the nice interface that Firefox provides is certainly a huge improvement. The article says that local caching of data will eventually erode the high number of search requests going to Google, thereby reducing the value of its search engine. Local caching makes sense for users who get their (locally available) content faster, can access cached content when off-line, and do not have to pay for bandwidth. It makes sense for the content providers that do not have to serve the same content multiple times to the same user, and it makes sense for ISPs whose (flat-rate) bandwidth is spared.

But before content can be cached locally, it needs to be searched and found (at least once) on the Internet...and there is a good chance Google will be employed to do this first search. Still, from my experience I will say that I often search the Internet for the same information. With a convenient locally accessible cache, I would cut the number of times I go to Google. Moreover, there is talk in academic circles about "leaf computing", the opposite of cloud computing. The idea of leaf computing is to build distributed computing platforms. For example, a bunch of like-minded individuals can self-organize their computers into a distributed, specialized web-crawler that goes and searches the Internet for relevant information to index, or syncs up the saved book-marks/local Firefox caches of several users. If there is enough bandwidth and storage on the 'leaf' computers, then why not move away from cloud computing and into a distributed leaf computing architecture?

From a Google Ad-sense point of view, I think Google may modify its search engine usage terms to have the right to serve ads even on locally cached content if this content was initially found using Google. Others may argue that all the value is delivered to a user the first time search results are used (like buying a song: you pay the same no matter how many times you listen), and so Google cannot serve ads for locally cached content. Does Google have the right to serve ads on locally stored content that it initially found? This is analogous to the TIVO advertisement issue of whether advertisement slots bought for a live show can be removed and other advertisements inserted for later recorded playback.

Sunshine in West Berlin

2009-03-09T10:46:00.002+01:00

This is why I love today!

View from the TEL building. 1 minute ago.

Wiring up my cellphone to my Web 2.0

2009-03-09T09:04:00.002+01:00

I have done 3 fun things over the past 2 months. First, I bought a very capable text.phone - the Nokia E71. Second, I connected my phone to my blog so I can send an email to my blog and it gets posted. And third, I set up a Twitter account where I can post messages via SMS.

Now this changes things. Suddenly I am untethered like never before. The ability to capture and publish information in real-time has moved from professional TV camera crews riding satellite TV vans to anyone with a phone that can send SMS.

That's change for the better. And one more example where tech and gadgets leave the nerd domain and change mainstream society. Information flowing freely, leaking through the porous gaps made possible through such emerging technology will loosen the grip of those who seek to control information, and by extension, the free will of people.

Blackberry's application store foray

2009-03-06T23:29:00.005+01:00

Blackberry is joining the app-store party by launching its own application store for the the Blackberry platform. There is a nice developer website and lots of buzz around RIM's latest move to take on the growing popularity of iTunes's app-store and the upcoming Nokia Ovi app-store. It is going to be an uphill battle catching up with the iPhone for all these me-too app-store ; but the Blackberry platform is a slightly different beast and its app-store positioning is quite different for some reasons.

First, most Blackberries are corporate property with significant organizational IT control over the devices. RIM may be positioning the developer tools and app-store for corporate IT development rather than for encouraging college-kid hackers who are trying to make a fast buck or two by writing a small game etc. This may explain why the Blackberry developer license ($200) is more expensive than the iTunes developer license ($99). In my opinion Blackberry isn't nearly looking to equal the number of iPhone applications. Its looking for serious business centric stuff on its app-store.

Second, Blackberries are not fun devices. When I see a Blackberry I start thinking of my consultant friends, complete with black suits and polished leather shoes! The Blackberry application consumer is going to be very unlike a 14-year old teen touting yet-another facebook widget on her iPhone. Instead, its going to be a corporate IT vice-president who likes an activity logging tool for the Blackberry to keep tabs on employees. Or an executive who downloads an extension of pocket-Excel for say, better readability. Bottomline: serious business applications.

Third, Blackberry apps will have to "work" much harder to gain the trust of potential downloaders. Blackberries carry confidential data and compromising this data could put the owners business (and/or job) at risk. Imagine the consequences of a software trojan that opens a connection to a server, dumps the contents of the Blackberry, and then blackmails the user or her organization? Certified and branded software applications have a much better chance of acceptance in the Blackberry user space.

Lastly, the Blackberry back-end is a unique add-on for developers. The "push" technology back-end of RIM can be used to create innovative applications on the Blackberry that may not be possible on other platforms. Question is, what is the (other) killer application for RIM push technology?

Social Network Regulation

2009-02-28T18:54:00.009+01:00

Telecommunication companies unbundled the local loop in the past decade to comply with government regulation designed to spur competition in broadband services. The regulation was enacted for allowing competitors to use the last-mile copper, which belongs to incumbent telecommunication companies, in return for a fixed (small) fee. For example, Germany has dozens of broadband providers that use Deutsche Telekom's infrastructure to reach all customers that DT reaches. A customer can switch from DT to another provider if she prefers. Therefore, the governments motive of creating competition in the marketplace has succeeded.

Now lets fast-forward to 2015. Social networking has become ubiquitious. The inevitable shakeout, M&A and standardization forces have linked up all social networks into one huge giant network. People communicate primarily over social networks. The search engine has become second to this the human Gaia - the social uber network. A collective virtual intelligence, call it society 2.0, overlaid over the society we know. Last time a technology, voice telephony, became the flagship mode of communication, government stepped in to regulate it, control it, and overhear if needed. Before last time another technology, snail mail post, became the flagship mode of communication, government stepped in to regulate it, control it, and read it if needed. And so it shall be with be with social networks.

How and when will social network regulation come about?

Reinventing the software development business for the mobile world

2009-02-24T06:13:00.013+01:00

Here is Pinch Media's presentation "AppStore Secrets", a data analysis of about 30m downloads of from the Apple iPhone application store. There were several insights in the presentation. The first thing that struck me was the relatively short lifespan of mobile applications compared to PC based applications.

Consider this: According to the presentation, most mobile applications seemed to "die-out" within days. They lie unused, or worse, they are uninstalled within a few days. Not even months. Compare that with Microsoft's MS Office, which usually lives on a PC for its entire lifetime. Usually a new MS Office user increases her usage over the months and years after purchasing it. While MS Office is expensive relative to mobile applications, the former's "per-run" cost may actually be lower than app-store mobile applications.

Fortunately the number of man-hours going into simple mobile applications for things like storing recipes and writing notes is quite small and most applications seem to be coming out of garage startups. I think this will change and we will see the emergence of a conglomeration of developers (e.g. a large mobile development company or even a developers' cooperative) who can amortize the risk and costs of mobile development. This may curtail rags-to-riches stories of 'lone star developers' as success will also be amortized. Still, the synergies of a large developer group working together, sharing code and risk make this a valuable preposition.

High level, reusable and modular tools that can be used to compose applications easily should emerge as the winners for mobile application prototyping and development. Developing everything in C++ from the ground up will not be economically feasible. Instead development tools like Python for the Symbian platform and .NET tools for Windows Mobile will become popular with developers. Given the variety of platforms, a cross-platform tool will have a huge advantage. In my opinion, Adobe's mobile development tools like Flash Lite are well positioned for such high-level and cross-platform mobile development.

iPhone AppStore Secrets - Pinch Media

Sirius XM satellite radio almost went bankrupt. A lesson for paid cellular mobile video?

2009-02-21T22:48:00.015+01:00

Satellite radio seemed like a good idea. Crystal clear, ad-free, broadcast distance-agnostic , high-quality audio content for those long commutes, all for about $10 per month. But apparently Sirius-XM, the 800-pound gorilla of US satellite radio, is in dire straits. It avoided Chapter 11 bankruptcy after Direct TV propped it up with funding, probably at the price of being taken over by Direct TV.

But what does this say about users' appetite to pay for content while they are on the move? If paid radio did not fly, then will paid mobile video not fly either? We have a simple analogy here

free FM : paid Satellite radio :: free Digital TV (e.g. DVB-H) : paid cellular mobile video

Its tempting to write up the requiem for paid cellular mobile video based on the satellite radio example. But here are some things going for cellular mobile video

Mobile VoD allows users to select what they want to see, unlike satellite radio, which is a broadcast medium.
Cellular video is truly untethered because it is consumed on the mobile device. On the other hand, satellite radio is not carried in the pocket - it is usually built into car entertainment systems. No extra equipment is needed for cellular mobile video unlike satellite radio.
Video is inherently a richer and more engaging medium than radio.
Telecom companies are going to sink a lot of marketing effort into making mobile video a success. Paid mobile video service is one of the most significant use-case for high speed wireless networks. And perhaps, the best escape pod from the dumb-bit-pipe scenario that telecom companies want to avoid.

Still, users' tepid response to satellite radio should worry cellular mobile video hopefuls.

Mobile application stores and Telcos

2009-02-20T00:09:00.009+01:00

Today's WSJ carries an article about online software application stores being the latest "me-too" business being promoted by mobile device manufacturers, telecom companies, and even independent companies. The first notable app-store came from Apple last year, for its iPhone and iPod Touch platform. Apparently IBM has already created a generic platform to create such app-stores, making it easy for any company to start its own app-store.

Third party mobile applications have been around for for a long time - ever since there were APIs and development platforms for mobile devices. I remember developing applications for a Palm III device using the Code Warrior IDE back in 2001. Microsoft also promoted Windows CE/PocketPC/Mobile development via Visual Studio and .NET from early on. In response the developer community created tons of mobile device applications that have been sold or given away for free on the Internet. So what has changed with the advent of the app-store? And why all this sudden interest? These are some of the obvious reasons for app-stores to exist:

The host company gets a cut of the revenue. For example Nokia will get a 30% cut on each sale in its Ovi software store. The developer gets to keep the balance. In addition, by providing this portal facility to developers host companies can attract developers who create cool applications, spurring the popularity of their devices.
The developer gets a suitable hosting platform for her applications. This includes a payment system, and users may be more willing to trust big names (Apple, Nokia) instead of a small unknown development company when they give out their credit card information.
The user gets a one-stop application shopping solution where she can compare applications and buy with confidence since hosting companies are most likely to certify applications for compatibility, quality, security, and legal compliance before selling them.

One of the challenges for telecom companies is going to be the consequent loss of control over the application mix on mobile devices and over mobile device hardware. Why? Because by creating an alternative revenue stream for device manufacturers, app-stores will force them to consider application developers' wishes while designing devices. Until now device manufacturers have largely depended on telecom-operated retail outlets to generate the bulk of device sales, making device manufacturers very pliable for telcos. Now device manufacturers can also lean on software developers, who create cool applications for their devices, to drive device sales and to generate revenue. The days when telcos could dictate device capability and decide which software got installed on mobiles are numbered.

Its not all red for telecom companies though. Smart phones and new applications are driving up data plan usage in a big way. There are some unknowns here, for example, users may start using flat-rate data plans for voice (through VOIP) and SMS (through IM and email) substitution. But will the cannibalization hurt telcos or will data plans compensate for the loss of voice and messaging revenue?

The Storage-Processing-Bandwidth triangle in mobiles

2009-02-18T23:33:00.001+01:00

The three fundamental "juices" of mobile devices - storage, CPU, and bandwidth - have been improving over the past decade. Storage has grown thanks to better and cheaper flash memory, CPUs have become spiffy thanks to Moore's law and advances in low-power CPU design, and bandwidth has improved thanks to 3G network roll outs. These 3 define the operating point of mobile device applications i.e., what are the trade-offs application designers make based on these underlying capabilities of a device. For example, if bandwidth was free and infinite then there would be less motivation to cache data on devices and instead most data (e.g. an address book) would be stored online. This would have the advantage of letting users access the same address book from different devices and computers.

But the truth is that while storage (bits per dollar) has doubled almost every year in the last decade, CPU and bandwidth have not kept the same pace. CPU power can still be jacked up if one is willing to trade battery life for faster CPUs. But bandwidth remains expensive (bits per dollar) and unreliable (spotty coverage) as compared to storage. Where is the operating point of upcoming mobile devices like smartphones heading? Will the distortion due to seemingly limitless storage lead to a redefinition of the mobile phone into a device that is less phone and more storage/CPU? Will voice and data lose their killer application status on the mobile device? What will users be doing with all that storage and CPU becoming available on their mobile phones?

It is safe to state that whenever data can be cached it will be cached. It just costs much more to transmit a bit than to store it, and transmission latency is also a problem. So expect applications to reduce bandwidth usage through local caching. High cellular bandwidth cost will also promote disruptive services like Wifi-enabled smartphones that can easily bypass operators' expensive data networks. With applications like Fring, users can already make phone calls using Skype on Wifi enabled phones. Mobile phone cameras routinely take images at multiple megapixel resolutions. But these high resolution pictures are seldom transmitted over the mobile wireless interface and are instead stored for later downloading to a PC. Many mobile device mapping solutions do not use cellular data networks and instead rely on storing all the maps locally. Multiplayer mobile games have not taken off either, partly because of the technical and pricing limitations of cellular bandwidth. I download my email over my home Wifi connection into my Nokia E71, reply to these emails on the train while commuting to work, and then connect via Wifi to send the replies once I am in office. Similarly, applications like Avantgo give users the option of avoiding the usage of expensive cellular data plans.

The argument I have tried to make is that the mobile device is not shackled to voice and cellular data only. Market forces will spur creativity to utilize these little wonders of modern technology in engaging and inexpensive ways. Voice and data may not remain the killer-apps on these devices.

Clearwire, Wimax, and cheap broadband

2009-02-17T21:44:00.000+01:00

There is a interesting article in the current Fortune about the troubles Clearwire is facing in rolling out Wimax on a large scale across the US. Apparently, it is working out to be much more expensive than was hoped.

In the city of Portland, Or, about 300 Wimax towers were needed to cover the whole area. Although in principle each Wimax tower covers several miles, uneven terrain can quickly spoil this. Clearwire offers Wimax for 30 dollars a month, a price that is significantly lower than wired broadband. But with the necessity of new hardware on the user side, will it work? Especially with Telcos supporting LTE as the Wimax alternative?

The answer lies on the business model for Wimax. Wimax has the edge in terms of early deployment and heavy weight supporters like Google and Intel. The key is the ubiquity of rollout as compared to telcos' LTE. Another question is whether Wimax equipment will be bundled as standard Laptop/phone equipment . Wired Broadband also bundles services like voice and increasingly, TV. How suitable is Wimax for these value-added services?

Test blog posting via email

2009-02-17T14:56:00.000+01:00

And this works!

The computing world meets the home electronics world

2009-02-15T10:35:00.012+01:00

One would think that computers and data networks can be easily integrated with home electronics like Hifi systems and TVs. Not so. Try hooking up a Hifi to your iTunes collection. Even if you are able to pipe the audio from the computer's sound card into the Hifi's line-in, its hardly convenient to browse and select songs on the computer and have them play on the Hifi. Computers are still lean-forward devices that require active user participation with keyboards and mice. I want to listen to music on my Hifi and browse my music collection sitting on my couch with a remote control. I want to control the music without having to unlocking the computer's screen saver and then using the mouse to control the software music player. And yes, I want to listen to my music with the fidelity that my Hifi offers.

Its interesting that content like movies, music, etc., produced primarily for home electronics, has moved from home electronics to computers and networks but the opposite movement of computing and networks features into home electronics has been slower. One could argue that computers can be easily programmed to mimic home electronics (e.g. software media and DVD players) whereas home electronics have to implement things in hardware - a more complex preposition. But modern electronics is easily capable of integrating computing and networking into products given how modular ASIC technology has become. Most home electronics has digital control circuitry since the 90s. These things are being controlled by computers for the past 2 decades! Then why aren't these computers interfacing to other computers and the big I(nternet)?

In my opinion, the reasons for this relative one-way street between computers and home electronics are more business and consumer related than technology related.

Silo thinking and protecting markets: For the home electronics industry, one living room = 1 Hifi + 1 TV + 1 DVD player + 1 Home theatre... Home electronics makers are reluctant to provide interfaces to connect up devices, especially digital interfaces like data networking capabilities, because there is scope to squeeze out redundancy. Today most living rooms have 3 audio amplifiers (TV, Hifi, home theatre) instead of one. But how long can this redundancy last if all music could be noiselessly (digitally) piped into one amplifier?
End users' lean back leanings: Home electronics buyers were people in their 30s-70s who were not too comfortable dealing with computers and Internet technologies. On the other hand users who used computers for entertainment were mostly younger. But an increasing number of people are totally at ease with the concepts of computing and data networks.
The computer vs. home electronics experience: There is certainly a difference between watching a DVD on a laptop and watching one on a DVD-TV setup. Up until recently this was enough to sway many toward home electronics. But the gap has narrowed significantly now. My wife's Dell XPS laptop comes with a remote control! Acquiring content is more convenient over the Internet (iTunes, Netflix, etc.) instead of acquiring it for home electronics (e.g.video store).
The other side: Unlike home electronics, the computer and telecommunications industry have been pushing any software application that consumes CPU, disk-space and network bandwidth - i.e. - music, video, and TV.

But things are changing. Our shiny new Denon M37 Hifi came with a USB port proudly placed on the front panel! The new war of the electronics world and the computer world. Or their confluence.

Smartphones and value for money

2009-01-13T06:37:00.003+01:00

Ok so I took the plunge. My shiny new Nokia E71 arrived last week. This is the first time I have bought a mobile phone at its retail price. Before this, I was usually happy to choose among the free phones supplied by the carriers. But this time things are different. There are fantastic smart-phones on the market (like the E71) and I have figured out the math behind why "free" is not free when it comes to mobile phone plans and 2 year lock-in periods.

Still, spending 300-odd Euros on a phone triggered introspection. Do I really need a smart phone that is, well, so smart. Smart is a relative term. I used to sync my old Sony-Ericsson to my PC Outlook and have my contacts and calendar up-to-date. That was smart too. So the question is, is my E 71's extra smartness worth the Euros? After 3 days, here are the things I am doing with my new phone that I could'nt earlier. Offcourse Nokia advertises the things you can do with this new phone, but I seriously doubt I will ever do all that Nokia suggests.

1. Do emails on the train.
2. Use the basic English-German dictionary in the phone
3. Type SMSes faster
4. Set up appointments, to-dos easily on the phone, thanks to its keyborad
5. Get an automatic weather report in the morning
6. Solitaire
7. GPS.. well, perhaps in the summer if I get lost or something (unlikely though)
8. I can chat and Skype-out using Fring.

Well lets suppose the phone lasts 2 years. that would get its price to under 50 Euro cents per day (discounting the cost of capital). That probably makes it worth the money given the list above.

...or am I rationalizing the splurge?

Wireless broadband options in India

2008-12-17T18:05:00.004+01:00

I am looking around for broadband services for my folks in Bangalore. They already have DSL via the state-owned BSNL. But this service is quite unreliable, being disconnected for days at a stretch. Before you draw conclusions about the India and her spotty infrastructure, be aware that this DSL broadband comes for about 5 Euros per month. That is 5 times lesser than comparable DSL in Germany. My way of looking at this is that part of the service offering (the reliability) is traded for the lower costs in India. Off course PPP (purchase power parity) will put 5 Euros in India right around 20 Euros in Berlin, but still, 5 Euros a month is a good deal - after all, people who buy broadband in India are very well off.

So we decided to shop around for cellular broadband services. The reasoning is that if you have a couple of services then at least one should be active at any time. The options are

1. Airtel (USB stick, 3G). Officially works at 115 kbps d/l, actually works at 15-20 kbps according to the salesperson, and costs 10 Euros a month for unlimited service.

2. Reliance (Antenna and coax cable, 3G). Apparently delivers upto 300 kbps, and costs 12 Euros a month for unlimited service. Requires installing an antenna on the rooftop.

I would go with option 2 for home use and option 1 for mobility. My folks have a better idea - just revert to dial-up when the DSL breaks because thats good enough for their need for checking email once a day. Cant argue with that!

Social networks - grade sheet

2008-12-09T19:18:00.002+01:00

I found a superb data analysis of the most popular social networking websites that may be interesting for some. The data was pulled from Google Insights and Google Ad Planner.

Look how Twitter is soaring!

Eric Maskin's talk at the Berlin/Brandenberg Academy of Sciences

2008-12-06T20:56:00.007+01:00

I had a unique opportunity to attend Eric Maskin's Nobel lecture (re-delivered) at the Berlin/Brandenberg Academy of Sciences. Eric Maskin shared the 2007 Nobel prize in Economic Science with Leonid Hurwicz and Roger Myerson for work in "Mechanism Design". This being my first brush with the topic, I expected a lecture of dense technicalities. I was mistaken.

Eric gave a great introduction to the theory and for once I started thinking of how great it would have been to take one of his classes! The idea of Mechanism Design is creating a set of rules that make selfish players behave in a way to achieve a certain outcome. In classical Game Theory the set of rules (constraints) are already given and then players try to maximize their utility (benefit) by playing by the rules via their self-chosen strategies. In the end, game theorists try to determine the steady state or Nash Equilibrium where no player can benefit by unilaterally changing his strategy.

Mechanism Design is about designing the constraints or set of rules. Eric spoke about things like writing the rules for green house emissions among the countries (the players). He also spoke about the energy sector and gave a simple example of how a mechanism can be designed to satisfy different players while making choices about which energy sources to develop.

Eric started with the brother-sister cake division example - how by letting the brother cut the cake in half and then letting the sister choose the first half the division would be fair. Mechanism Design - my brother and I(re)invented this one when we were 6 ;-) !

The rise of the Netbook

2008-11-30T15:58:00.012+01:00

While walking in Media Markt's laptop aisle yesterday I was surprised to see the number of mini laptops (Asus EEPC, Acer Aspire 1, Toshiba netbook etc.) in the €299-500 price range. Interestingly, some of them are being sold like cellphones - with a price tag of just 1 Euro with a 2 year "3G data plan". The data plan costs 37 Euros, and probably gives customers a few GB of bandwidth a month.

So adopters get to carry their *free* netbook and Internet connection anywhere, without paying outrageous Wifi hotspot/hotel Internet charges. Perhaps in a few months this Netbook offer will spur many to switch to 3G Internet and give up on tethered Internet (cable, DSL) entirely. Are we looking at this cannibalization in the next couple of years? Low cost cellphones cannibalized fixed line telephone users. Can netbooks with 3G data plans cannibalize tethered DSL?

It all depends on the service quality of 3G data vs. DSL. Now 3G data can seldom serve more than 100s of kbps versus the multiple mbps of DSL connections. But I hazard that there is a sizeable market segment to whom the mobility (and free netbook) will appeal more than blitzy Internet connections.

Another issue will be the scalability of 3G data - 3G infrastructure has some fundamental bandwidth limits - which make a mass deployment in dense areas problematic. There is LTE, the next generation of cellular wireless networks that promises much more bandwidth, but its deployment is only planned over the next decade. Perhaps mass data demand over wireless will speed things up for LTE (or WiMAX).

The multi-person home tethered service is probably safe from wireless 3G data for now. Splitting 3G bandwidth is certainly possible via Internet sharing or a router that accepts a 3G data card and then wifis the bandwidth to multiple home PCs, but the slow speed will be an issue for demanding applications like online video, gaming, etc. 3G data is still a 'midband' service. Moreover, netbooks are difficult to use with their small screens and cramped keyboards. Presently deployed laptops and desktops are certainly more comfortable (minus their limited mobility).

The Netbooks' rise is absolutely remarkable. A couple of years ago MIT's OLPC project was the seed of the idea to create cheap affordable laptops for school children in developing countries. Then Intel threw in its own school laptop competitor. Asus stole the commercial show with its EEPC 701. Intel's spectacular strategy of creating the low-power Atom processor created a captive netbook market for Intel and allowed for smaller and lighter batteries. Microsoft resurrected Windows XP as the Netbook operating system while Linux provided a free alternative to Windows XP. And then there is the 3G data availability through USB sticks.

It all seems to have come together at the right time.

Pure gold statistics about the Internet

2008-11-08T20:19:00.008+01:00

Figure: Internet users by age (Click to enlarge). Reproduced from here

I found an absolutely remarkable report issued by the OECD in the summer of 2008. It has got lots of goodies on OECD Internet usage. Here is the URL

The future of the Internet Economy, a Statistical profile

One graph in the report that was very interesting and it is pasted above. The graph shows OECD countries' Internet usage conditioned on the age-group. It is clear that younger people use the Internet much more than older people.

Now as time progresses the young people will age, and I do not think the young people of today will give up on the Internet as they age. Meanwhile the next generations will be even more Internet savvy than today's youngsters. So the low down is that the next 15-20 years will see a high growth period for OECD Internet broadband demand. There is about 20-30 more years of growth in this space before Internet broadband growth saturates, say, the way Electricity did.

My splintered digital world

2008-10-29T13:29:00.004+01:00

I get the feeling that my digital life is divided, unequally, among my email accounts, my phone, my USB stick, my laptop, my desktop, my cameras, my linux workstation, my disks, my backups, my Facebook, my Linked in, my blog, my webpage, my o my o my.

When I first started using the PCs 2 decades ago I had a nice floppy disk to carry my digital data. It fit in, or it was thrown out - great garbage removers those 360kB DSDD disks. Well I am not going to go down the Luddite path here, I am perfectly happy with todays infinite storage. But I am concerned about data splintering. My data is in all these various formats, versions, and names somewhere in various parts of my digital ecospace.

Now not to conclude that I don't have a system to file different things in different directories and the like. But too much information may result in suboptimal storage. For example, I have nicely named and dated folders for my pictures. But when 94 photos of my niece arrive via email it takes precious minutes to download and store each picture. How do I aggregate information quickly and without manual labor?

Algorithmic search presents the next best alternative - just keep everything anywhere and then have your computers crawl abd index the information. But search does not span devices (At least right now). How do I pull up the phone number stored in my home phone's caller ID while sitting at work?

Problems problems problems. Thats great because this means there is a whole lot of work to do in this area. Start-up anyone?

Smartphones or flash drives to replace the laptop (?)

2008-10-27T07:43:00.009+01:00

I am intrigued by the possibility to leave my laptop at home (or work) instead of carrying it around every single day, as discussed in this WSJ article. Are we already there?

I have been running a small experiment on myself for the past few months about the feasibility of this approach. My use case is one with heavy usage of my Outlook mailbox, lots of documents, and some software like MS Office, Emacs and Matlab.

In my experiment I have stored all my working data files in an 8 GB Sandisk cruzer USB drive. I plug the cruzer drive into various computers I have access to, just as I would a smartphone with so much flash memory. It mostly works (i.e. I don't miss my laptop), but here are the unresolved issues

1. Security: Yes I mean the consequence of loosing the flash drive (smartphone), but also the issue with secure corporate Outlook email on my laptop via VPN and certificates. It is impossible to have the same corporate setup at home on another computer (at least where I work). But this may not be an issue for those who use web mail.

2. Software: Well lets face it, not all software can be installed on every computer. The other option is trying to install software on the flash disk, but then many software installations bind themselves to the computer - for example - those registry keys of MS Office installations. Perhaps this is an area where more innovation is needed to untether software from hardware. For now, I use my computer agnostic Emacs editor as my data input tool out side of my laptop. Oh, and I also use it when I am on my laptop. I love it!

3. Customization: There are ways to copy your browser favorites, screen savers, wall papers etc. on your flash drive or smartphone, but I would say figuring all this out is cumbersome. Instead there is this cool Mokafive concept of carrying your whole OS and data and customized software all on one flash drive! Just boot off the USB drive and you are done. I found out from the IT guys though that security software will complain about this. Another problem is that loading and running an OS off the USB drive will be slowwwwwwwww.

Lala, streaming for life, cloud computing, bandwidth, and the ISP

2008-10-22T18:00:00.006+02:00

Lala says it will sell you the right to stream a song, for life, for 10 cents. Thats quite a sweet deal for someone who listens to music on the computer only. In case you really want to take the song on the go, you can buy the song permanently for your music player for the same 99 cents. Lets do some bandwidth Math now.

Use case 1: Fixed line user
A user streams 8 hours of music every day @ 128 kbps from Lala. First off, 8 hours at 4 minutes per song is about 60*8/4 = 120 songs. If the user's Lala library has a different 120 songs for each day of the week, s/he has 600 songs (= 120 * 5). An investment of $60, for a lifetime. Now this is quite a good deal compared to the corresponding $600 based on the current 99 cents-a-song model. Offcourse you loose the right to download the song into your iPod, but for this use case lets say it doesnt matter to the user. The critical point is, users will not download the song one time as in the current model but will download it everytime they want to hear it.

Now lets do the bandwidth calculation.

(8 * 3600) seconds * 128 kbps = 460MB

So we have a 460 MB sustained streaming download per day. For 20 week-days a month, we are talking about a bandwidth usage of 9GB per month. This usage certainly puts use case 1 into the ISPs' "power user" category. Do we have enough bandwidth provisioning in the core and access networks to deal with large numbers of such users?

Use case 2: Mobile Internet user
Use case 2 is a mobile internet user (think UMTS on a laptop) user. Even if we cut the music streaming about 1 hour per day, we have a usuage of over 1GB per month just for music. Do we have that sort of bandwidth on 3G networks and will flat-rate data plans tolerate such perfectly legal users?

Van Goghs from the S&P Stock Index

2008-10-19T16:59:00.017+02:00

(Figure 1: Click to Enlarge)

I was playing around with the S&P historical data (monthly averages from 1871 to 2008) and came up with Figure 1. In this figure I show the value of $100 invested at each month since 1871 in an S&P index fund (see my related post) and this is the first independent axis. Another independent axis varies the lead time to sell, i.e., the time the investor waits for before selling the invested fund. Finally the dependent (z, vertical) axis shows the total return (principal + profit/loss) on the $100 that was invested initially.

This 3-D graph is in itself quite interesting although it is too dense to offer any direct insights. So I flew to the top of the graph (the virtual geek way - I set the viewing azimuth to 0 and the elevation to 90 degrees). And then I created my Van Goghs of the S&P Stock Index!

Figures 2a (24 months) , 2b (5 years), and 2c (10 years)
Click to enlarge

In Figures 2a, 2b, and 2c, the vertical axis is the time of making the investment of $100 in the S&P index fund. Blue signifies losses, and hotter colors (reds, yellows) signify profits in the color maps - notice that Matlab has assigned different colormap scales to each of the figures.

The horizontal axis is the time for which the investor holds on to the index fund before selling it (in years). Note the dark blue streak around the Great Depression (1929-) in all the graphs. It gets thinner as you move from right to left - since someone who exited just before the big fall saved themselves, but those who had invested earlier but held on lost (blues). You see blue lines around the year 2000 - when the Internet stock bubble burst. But you also see the dark red streaks of pure profits interspersed throughout the graphs.

Search for the rare combinations of small lead times and large profits in the 3 figures. Thats where investors invested and were quickly able to make large profits - if they exited wisely. And that will make for a wistful "if only I had invested and divested in those red streak times!"

Impressionist no doubt!

Why to (still) believe in the Stock market

2008-10-15T16:29:00.005+02:00

( Click to enlarge)

One of the most exciting things about the stock market is its unpredictability. Some liken it to casino gambling in that there is randomness in the return on investment. Moreover, the conventional thinking is that the odds are stacked against the small invester since he is competing against highly organized hedgefunds and mutual funds with talented fund managers. So the question is, can the ordinary investor make money on the stock market. I know the answer is yes, but by "ordinary investor" I mean someone who just uses a simple mechanism of buying stocks at a low price and the selling them after a certain time lag when prices are high. Nothing fancy like short selling, derivatives, etc.

I did some basic analysis on S&P historical data to debunk the first misgiving about the stock market being a casino, and events for the last couple of weeks (Oct. 2008) have debunked the myth of the know-all big fund. They are all bleeding red ink as much as small investors (no, portfolio diversity didnt save the day for them, but that is for another blog post).

Lets get back to the S&P historical data. I obtained the monthly S&P averages from the January 1900 till May 2008. I then wrote a Matlab script to invest in an S&P index "fund" during each month an amount of $100, and sell this after a pre-specified lag (1 month, 12 months, 5 years, 10 years). The figures for the different lags are given above (click to enlarge) and show sale value (Y axis) of the $100 investment that was made in the S&P index in the time specified on the X axis. These graphs are not adjusted for inflation.

For small lags (1 month, 12 months), the graphs look random and seem to support the casino effect. However, for larger lags (60 months - 5 years, and 120 months - 10 years) there is a clear trend. Some times were better investment times than other times. For example, it was smart to buy in the late 80s when the stocks were low and sell during the late 90s when the stocks were high (120 month lag).

Some conclusions from this basic analysis are

There is room for applying basic intelligence, and hence, this is no casino play where winning follows a certain probability distribution.
Timing is everything when considering stocks as an investment. It is as important to guess the selling time as it is the buying time. For example, folks who bought in 1920 did very well by selling in 1925 rather than 1930. Buying and then keeping stocks away like fixed-term treasury certificates is a bad idea.
Short term gains in index funds are hard to come by. Try specific stocks for this (and assume the greater risk of no diversity in this case).

I am still working on this analysis. Will keep this blog posted If you want the Matlab scripts just email me.

Where is all that Wall Street money?

2008-09-27T08:21:00.010+02:00

You've all heard the news - investment houses going bankrupt, being sold off to retail banks, large insurance companies being nationalized - all within the span of a couple of weeks. What amazes me the most is the sort of numbers floating around. A year ago the market capitalization of these busted companies was 100s of billions of dollars. They employed 10s of thousands of people, including ivy league-educated finance jocks and MBAs. Their stock prices seemed to go up, up and away - a complete endorsement of their magical money making ability. Then, ignominiously, they went broke.

Monetary circulation is a closed entity - dollars don't just float away into outer space. So I am thinking - where did all those loaned dollars go? To understand that, lets look at Mr. John Doe's 4-bed/3-bath home in suburban San Diego that was built ca. 2004 and bought by Mr. Doe at a hefty price via a loan. Say Mr. Doe has fallen behind on his payments in 2008, therefore adding to the toxicity of the CDOs - Collateralized Debt Obligations - that wrongly counted Mr. Doe's mortgage as AAA+ reliable. But all this happened in 2008. Where did the money go to in 2004?

The developer bought the land from the state. Therefore a part of the loan capital went to the state. The house itself was built using superior building materials (expensive house) and therefore, part of that capital flowed into the pockets of the building material company shareholders - the glass company, the wood company, the lighting company etc. A big part of the house price was profit for the builder/architect company and therefore, it went to these companies' shareholders. There was Latin American labor to built the house, and so some of the money went to Latin America via Western Union transfers. Some more must have flowed to China for building materials, or perhaps to Italy for the Italian marble.

Now the key point is that the value of the asset handed to Mr. Doe was supposed to rise as time went by, because this house was in the San Diego area, with the beautiful Southern California climate, the wonderful, peaceful, and happy society, the good public school in the neighborhood, and consequently the never ending demand for housing as people from all over the world came looking for a piece of this beautiful part of the world. In fact, Mr. Doe bought the house factoring all this into the future equation to pay back the hefty mortgage. In the worst case (he thought), he could just sell the house and pay back the mortgage, making a neat sum for himself. And until he sold, he could live a good life in the expensive home.

Unfortunately for Mr. Doe (and everyone else), the price of his house actually fell, and this voided the whole argument of the previous paragraph. Now if Mr. Doe's house goes into foreclosure, Mr. Doe's lending bank will only recover the reduced price of the house. The notional and fluffy value described in the previous paragraph could not be converted back into hard money when it was needed in 2008. Money has been lost, and this fact bubbles up to all those CDOs on Wall Street. Until the value of the asset -that house - rises again, there is no way to fix the problem.

Hype around "Hypertargeted" advertising, and what REALLY matters in click-based advertising

2008-08-04T20:11:00.011+02:00

This article in today's Wall Street Journal discusses Myspace's "Hyper targeted" advertising system. The system studies profiles, messages, and other information of Myspace users and divides users into more than 1000 distinct "buckets" or categories. This classification can be used to target specific customer groups very effectively. Or so News Corp. (which bought Myspace for about $580m) hopes.

The article goes on to explain that "hypertargeting" has had checkered success, with some advertising campaigns having moderate success while others doing perfectly well with less "targeted" and more generic location-based (zip code) online advertising.

There was one very interesting example in the article. Quoting from the article

"The New York Health & Racquet Club spent $5,000 on a MySpace campaign that displayed 2.3 million ads to users on the site. Though the health club could have chosen to target ads at people who say in their profiles that they enjoy rock climbing, yoga or working out, it chose instead to simply target by age and ZIP codes near its facilities. The club said it was relatively happy with the campaign, which generated roughly 1,000 clicks, a response rate of just 0.04%."

-Source: WSJ

Now lets see, the New York Health & Racquet Club spent $5000 for 1000 clicks, i.e., $5 per click. If we assume that 5% of those folks who clicked on the ad actually signed up for club membership (means 50 sign-ups) then the per-membership marketing cost is $100 per customer. Not bad, considering that the average membership is $75-$100 per month. One the other hand, if only 10 people signed up, then you have a much higher price of $500 per new customer.

The key question therefore is, what is the post-click conversion rate, i.e., yield per click. Because this determines the value of the click for an advertiser (like the health club), and by extension, the price of the click that a content syndicator like Myspace can set. Ultimately it is the advertiser's landing website that needs to make customers out of users.

The Myspaces and Googles of the online world may well find it worth their while to start helping advertisers convert clicks into dollars instead of stopping at matching the exact user profile with the exact advertiser. Hypertargeting is good, but paying customers are much much better.

What matters to me, whats in my head, and this blog

2008-07-22T22:14:00.004+02:00

Figure: My Wordle view (Click to enlarge)

I came across Wordle - a service that lets you create a word cloud from any text, highlighting those words that occur frequently. The above figure is generated from the text of this blog. The picture says it all!!!

Tesla, EVs, and their mass adoption

2008-07-20T18:42:00.020+02:00

Figure: Tesla Roadster, the sporty Electric Vehicle

The current Fortune has an article about the teething troubles of the Tesla Roadster, an electric vehicle (EV) being touted as an all-electric sports car (click here for more pictures). Apparently, more than a 1000 people, including some who-is-who s, have signed up to take delivery of the first fully electric sports car. The article states that excitement remains high, never mind that Tesla is having problems keeping the delivery date for most orders.

Tesla's website says that the Lithium ion cell powered vehicle can cover 220 miles per recharge. Now that is quite impressive, if you consider that according to the AAA an average American drives only 29 miles per day. As long as you are not driving cross-country the Tesla Roadster should almost replace your conventional sports car, almost because recharging the 6,831 Lithium ion cells on the Tesla Roadster takes 3.5 hours as compared to the 5 minutes of tanking-up the conventional Porche Boxster. The long recharge time is still not a deal breaker - if you can remember to charge your cell-phone every night then plugging in the car every evening shouldn't be that hard either.

The question is, will EV technology follow the conventional wisdom that early adopter products migrate down to the mass market? Does it make economic sense to buy such a car for the John Doe on the street, if not now, then 5 years into the future?

The Tesla motors website says that the operating cost for the Tesla Roadster is under 2 cents per mile. The operating cost for a comparable Porsche Boxster is about 20 cents per mile (calculated from this website, with gas at $4 per gallon)*.

Unfortunately, the Tesla Roadster 2009 edition costs about $109,000 while the Boxster costs less than half, about $50,000. Or, put another way, you will have to drive

(109000-50000)/(0.20-0.02) = 327,778 miles,

before the extra price of the Tesla Roadster can be justified!!!

Since Lithium ion batteries will not last 327K miles (neither will the rest of the car), I think that EV technology is not getting into the mass market anytime soon. Even if the price of gas triples, you will have to drive more than 100K miles in your EV before it saves you any money. And I haven't even factored in the lost opportunity of investing the $59,000 difference elsewhere.

So clearly the argument of saving on energy costs is meaningless if the EV is going to cost an arm and a leg. Question is, can EV manufacturers, or liberal government subsidies, narrow the price gap between EVs and gas-powered vehicles?

*assuming that the operating costs only cover energy costs

Crude oil: how much do we have?

2008-07-13T17:59:00.007+02:00

Figure 1: World proved crude oil reserves (Billion Barrels), click to enlarge

Its always good to see how much ice-cream still remains in the tub in the freezer. So I visualized the World's proved crude oil reserves (Figure 1) with data from the Energy Information Administration. The EIA provides a wonderful excel sheet with all this data, I just made the plot.

I like some things here:

The discontinuities - sharp jumps - upwards.
The fact that the world-wide proved crude oil reserves have more than doubled in about 3 decades.
Most of this doubling happened in active oil producing regions (the optimist in me thinks that more prospecting in other under-studied regions may yield some more discontinuities, in the right direction, i.e., up).
We seem to be pumping out less than we are discovering (thats why the aggregate proven reserves point upwards)

So, why are crude oil prices shooting through the roof if there is so much buried under us? These are some supply-side* reasons:

Crude oil is harder to get because new reserves are geographically challenging.
Sweet light crude is harder to find, and oil companies need to look at harder-to-extract and harder-to-refine heavy crude.
There is not enough refining capacity.
There is not enough investment in new oil fields.
Some of the crude oil lies in politically unstable regions.

Still, we have a whole lot of crude oil left. My bet is that during my lifetime (next 4-5 decades) we will end up with more proved crude oil reserves than today due to new exploration finds, and also because demand will drop due to the World switching over to alternative fuels.

Amen to my optimism!

*For some demand-side analysis see this post.

Unlearning Google

2008-07-05T10:39:00.016+02:00

Figure 2: Light tailed website usage (Click to enlarge)

Figure 1: Google vs. Non-Google properties in my Firefox history (Click to enlarge)

Google's Viacom fiasco is an ominous wake-up call for anyone who cares about his or her online privacy. Today Viacom, tomorrow some other company, another day a government, can arm-twist Google into giving away log data containing user names, IP addresses, keywords, watched content, mouse-clicks, email, and any other information that Google collects.

So far, Google has only used user data for directed marketing. At least it is only about wringing money out of people's thoughts and desires through the ad sense infrastructure. The problem is, the same data can be easily massaged into revealing political, ethical, racial, religious, sexual, and other personal leanings of a person. There may be money to be made out of this data as well, but more importantly, there is the real danger of misusing this information as a pretext for prosecution or blackmail.

Google publicly defends its privacy record. Unfortunately, user privacy is not the most important objective for a publicly traded company. It is shareholder value. And to create shareholder value, a company needs to survive. A determined government can easily make the survival of a company subject to compliance with the government's wish. Google says it "Does no evil". Trouble with this slogan is, who decides what "evil" is?

Another scary scenario can be built around theft of sensitive user data. The media reported that Google is handing over 4TB of You-tube log data to Viacom. Now 4TB is a substantial, but not a lot for future data storage technology: We may have 4TB USB pen drives within the next 5 years. What if one disgruntled employee smuggled this data out of Google and auctioned it off to blackmailers for a few hundred grand?

No easy answers here.

I can keep ranting about Google and privacy and all that, but I am writing this blog on Google property (Blogger)!!! My wife and I are avid Gmail and Orkut and Google Reader and Google search and Google news users. Are we toast? Or, can we wean ourselves from Google?

I parsed our Firefox history over a few weeks to figure out where we stand in terms of Google-to-non-Google websites visited in order to get an idea of our Google dependence. The results are not pretty. Google properties accounted for just over 50% of all the websites visited (Figure 1).

Fortunately, there are non-Google alternatives to all Google applications. So in theory we can start using other applications instead of Google. Off course, there is nothing to guarantee that other websites will not yield to the same pressures as Google. But at least we can spread our web footprint - one entity will not have a complete view of a our web presence as Google does today.

The Firefox history indicated that we visit a few websites often and the rest are rarely visited (Figure 2). The often-visited websites were the usual suspects - search, web-mail, social networking, blogs, and news - and Google dominated this space. This is a great sign because it shows that even though Google is big in terms of visits, it is not very heterogeneous in the content/services it offers. Google is not my bank, not my bookstore, not my voip provider, not my university, and not my community. In fact, if I remove the top-6 Google properties from the data then the distribution starts looking much more uniform. My web log data spread on heterogeneous websites. Doesn't this flavor of obfuscation help privacy?

There may still be hope for privacy on the Internet.

Hollywood DVD - a buck a pop, beat that...maybe with ice cream?

2008-06-29T18:59:00.014+02:00

My primary source of Hollywood entertainment is the Videocenter movie rental store located at a stones throw from my flat in Prenzlauerberg. Each movie is available for 1 Euro per day, the collection is as current as the DVD release schedule, and the place is run by courteous bunch of folks. There is a popcorn machine, soda fridge, snack isle, and even the Ben & Jerry's cooler. All yours, for 1 Euro. 1 movie per weekend * 1 euro per movie = 4 Euros per month.
Thats the monthly Hollywood bill for me and the wife. Now that is a hard-to-beat deal.

If I was wearing a cable/IPTV VoD service provider's hat I would be hard-pressed to beat this deal because breaking even at 1 Euro for another distribution medium is a tough cookie (see my related post on Netflix's VoD distribution cost). Plus cable/IPTV cannot deliver my Ben & Jerry's ice cream tub.

Wait, that gives me an idea. Maybe Cable/IPTV VoD service providers can team up with ice cream trucks to have them deliver ice-cream and get a cut from Ben & Jerrys. Pizza, snacks, popcorn, T-Shirts, I don't know, movie specific stuff. Perhaps this may allow service providers to compete with Videocenters. The thing is, they need to start looking outside technology and into ice cream trucks.

Renewable energy : Fossil fuels :: David : Goliath

2008-06-22T21:32:00.028+02:00

Renewable energy annual growth scenarios (Click to enlarge)

US annual energy flow (Click to Enlarge)

I was reading the latest Economist (June 21st-27th) which talks about the next innovation boom for VCs to invest in: Energy. The scale of the Energy market ($6 trillion, about 10% of the global economy) is at least one magnitude higher than Internet and Computing (booms of the 80s and 90s) put together. No wonder that Energy is generating so much interest in the investor community and among researchers and entrepreneurs.

The driving force for innovation in alternative energy sources like wind energy, solar power, and bio-fuels is the steep increase in crude oil and natural gas futures (No I don't believe it is out of love for the environment). I wanted to understand how much time is needed make a significant hole in the fossil fuel demand by way of diverting the energy demand to alternative fuels. A beautiful figure from the 2006 Annual Energy Review released by the US Energy Information Administration is a nice starting point. The figure is US specific, and does not consider energy hungry China or India, but if the US energy juggernaut can be tamed with say, 50% alternative energy sources, then I am certain that China and India will happily adopt these viable alternative energy sources as well. (Plus I don't have the beautiful figures for the rest of the world so lets work with the US data!)

About 14% of the total US energy comes from renewable sources and nuclear power, in fact the figure also says that only 6% of the energy comes from renewable sources excluding nuclear energy. The rest comes from fossil fuels (including natural gas). So lets try to guesstimate, based on this data and varying rates of renewable energy growth in the coming years, the time until we derive as much energy from renewable energy as from do from fossil fuels today.

I have plotted 4 scenarios based on 5%, 10%, 15%, and 20% annual growth of renewable energy starting from their base 2006 value (from the Energy flow diagram). The plot indicates that it is going to take between a 10% to 15% annual growth of renewable energy in order to catch up with the present fossil fuel energy contribution by 2030. While there is no hard-written reason for the annual growth to not exceed 10-15%, I believe that there are significant inertial factors, like deployed fossil-fuel based equipment, lack of skilled engineers, innovation lag, legal issues, etc., which will keep renewable energy from growing at higher annual rates.

Conclusions:
Basically, this simple back-of-the-envelope calculation indicates that we need sustained double digit growth in renewable energy over the next 2 decades to challenge the fossil fuel Goliath. The figures also seems indicate that in the short term (5-10 years) fossil fuels are going to be the primary energy source. Therefore the world is going to need either a huge increase in supply or an appreciable decrease in demand of fossil fuels notwithstanding any alternative sources of energy in the short term.

Supply will grow as better technology is used for oil and gas exploration. It has already become worthwhile to use high-sulfur crude instead of sweet light crude oil. But the fact remains that demand will have to abate to meet the short supply through higher prices and unfortunately, slower economies.

Net neutrality: The value of a byte travelling on the Internet

2008-06-15T21:53:00.013+02:00

In response to Richard Bannet's article.

Let us say, for simplicity, that most of us are connected to the best-effort, statistically multiplexed, Internet. What does this mean? This means that every byte on the Internet going from point A to point B will, on average, get the same service from the Internet (same probability of loss, same delay, same delay-jitter, etc.) Therefore the Internet has the tendency to treat each byte traversing it as equal because in our simple example of 2 bytes going from A to B, the fraction of service (or utility) that each byte receives from the Internet is equal.

However, most people agree that the importance, or utility, of every byte on the Internet is not equal. For example, it may be more important to quickly transfer a byte from a voip conversation than a byte from a file transfer. Or it may be more important to send bytes that update stock prices than to send bytes to play a You tube video.

Or so I think. But what do you think? What does Skype think? What does Google think? What does Comcast think? What does the government of a country think? And if they think differently, then whose voice matters? Or should anyone's voice matter more than the others?

This is the key point of the Net Neutrality corundum. Everyone agrees that the present design of the best-effort Internet is suboptimal in that it treats every byte as equal and gives equal precedence to equal fractions of content. But the issue with doing away with this Net Neutrality model is that vested interests will decide which particular byte is more important than another byte. Can we trust one single company, or authority, to make the correct decision on this one? As a market believer, I would first say that let the market decide, i.e., let the price per byte, and hence the value attached to that particular byte, be the deciding factor. But the big issue is whether a flexible, effective, and dynamic market of this sort can be set up and quickly integrated with the existing and upcoming Internet protocols. Until this happens, I am more comfortable with time-tested simple statistical multiplexing, the fair but sub-optimal egalitarian algorithm, to do the job.

I am relieved that the question of Net Neutrality does have a technical solution - setup a market to do the job. I am just concerned of whether there is enough political patience to wait for the technology to develop this byte market we will need.

Scott McNealy's talk at TU Berlin

2008-06-11T16:40:00.033+02:00

Scott McNealy at TU Berlin

Scott McNealy gave a talk on "Open Wins: Leadership and Innovation" at the Technical University of Berlin today. Scott McNealy is of the "The Network is the Computer" fame from the dotcom boom of the 90s and is presently the Chairman of Sun Microsystems. I was part of the audience and found a few things worth blogging.

The key message of his talk was that Sun is and has been an open source champion forever. I don't buy into that (Solaris wasn't open source till 2005). But Scott blamed a prior agreement with AT&T Unix for this. Whatever the truth is, I cannot seem to place Solaris in the same bandwagon as Linux when I think open source. I suspect that Sun is trying to leverage the open source developer community to shoulder the costs of keeping Solaris updated (he himself alluded to the bug-squelching power of open source software). Perhaps like the Fedora project of Redhat?

Scott also touted the Sun Ultrasparc T2 processor - with lots of multi threading and multi core support and the 1.5 Watt/thread low-power footprint. Off course there was the usual Microsoft/Oracle bashing (Microsoft and its patches... so un-open source, Oracle and its $40k per-core licensing... so un-Mysql), but the hidden message was that Sun hopes that these new T2s find their way into routers and energy conscious data-centers. Although in my opinion I doubt if Oracle's grip on enterprise database computing is going to loosen up anytime soon, so Mr. McNealy may find it a tad-bit difficult to wean enterprises from their established Oracle databases into to Sun-acquired Mysql.

I asked Mr. McNealy about why Sun-promoted Open Office does not enjoy the MS Office - .NET (think VB Macros) type of integration with Java. The answer I got seemed to suggest that I should ask the open source community about this. Why shouldn't Sun take the lead in this? Java is their baby and Sun says it is all for Open office, then why not cure the Achilles heel of Open office through Java integration? This makes me suspect what I said about Solaris earlier - that Sun is trying to leverage the open source developer community to shoulder the costs of integrating a credible scripting platform into Open office.

Undoubtedly Sun has done more for open source than most other companies (for example, as Scott rightly pointed out, Google has a rather poor record on this front). Lets hope Sun does a lot more in the future.

Entertainment technology and diminishing returns

2008-06-07T19:45:00.014+02:00

I am trained in information theory, a science which emphasizes squeezing every last bit of efficiency out of communication channels to get a voice or video signal across in its best form. A noble endeavor no doubt, and one that has spawned a whole analog, and later digital, entertainment industry. But how much does it matter to the end-user?

My basic thesis is this: up to a certain quality people do care about how much error-free information gets across. But after that the human brain's smoothening kicks in - the apparatus in us that can skillfully ignore any small blemishes in the audio track or on the screen. While it is true that video technology already takes advantage of this "help" from the human brain (thats why finite frame rates and digitized pictures work), I have a feeling that sometimes technology needlessly pushes bits which are not useful, leading to "diminishing returns" on entertainment technology investment. After all, how many of us can make out the difference between a 192kbps-encoded MP3 file and its uncompressed counterpart that is 10 times bigger?

For another example, lets look at HDTV. Unless you are watching from close distance, the low pass filter in your eyes will substantially smooth out the sharp images on the HDTV screen. Undoubtedly HDTV looks better (but how better?) than SD TV, but is the delta enough to drive consumer pull in the mass market?

The rapid sales of HDTV s and blue-ray disks seems to suggest so. But I'd like to know what fraction of the content being seen on these HDTV s is really HD? And when the world does shift to HD, will the lowly SD TV be forgotten, going the B&W TV way? Probably not, because a vast library of content is stored in SD TV format. My kids will probably watch my old Friends and M*A*S*H DVDs or my father's music video collection (stored on VHS!). So their eyes and senses will probably accept fuzzy-ol' SD TV as well. Entertainment is about content quality first and then about technology quality.

Wildflowers in the Prenzlauerberg spring

2008-05-25T19:58:00.014+02:00

(Click to enlarge)

May 25th, 2008. A gorgeous afternoon in Prenzlauerberg, Berlin.

Me and my Nikon D40 were out to capture the beautiful wild flowers of Prenzlauerberg. Here are a few samples from the excursion.

Finally, the "Net" in Netflix; plus, the bandwidth question

2008-05-21T17:19:00.026+02:00

Netflix has released a set-top box that users can use to receive movies directly over their broadband Internet connections. The box, developed by the silicon valley company Roku, has received good reviews on CNET and PC Magazine for its nice interface and more-or-less good performance over most home-broadband connections.

Advantages for users

No propagation delay from snail-mail shipping DVDs - no more waiting for 2 days.
No need to mail back DVDs.
Ability to switch to another movie or show - you are not stuck with that wrong movie you placed in your Netflix queue.
No extra cost except the broadband connection and the $99 cost of the box.

Advantages for Netflix

Savings in storage, handling, and shipping costs (to-and-fro) of the DVDs. Theoretically, if all Netflix subscribers switch to this technology then Netflix can close its nation-wide distribution centers and also save on postage costs: assuming that Netflix pays the standard first-class mail rate of $0.42 to USPS, thats a $0.84 saving per mailed DVD. I think that the present overall cost of circulating a DVD to a user may be well above a dollar for Netflix.
Centralized content control and the ability to speedily deploy new movies, shows etc.
Ability to expand beyond the US in a relatively painless way - no distribution centers to set up, no additional staffing costs (analogous to how iTunes operates in Europe).

And the Bandwidth cost?

The Netflix system delivers video streams at 2.2 Mbps, 1Mbps, and an even lower bit-rate depending on the connection between the server and the receiving box. The quality naturally degrades according to the lessening bitrate, but let us assume that a user has a great Internet connection and that no bandwidth bottleneck exists between the serving CDN and this user and so s/he can watch the best 2.2Mbps quality for the entire 120 minutes of a movie.

Size of the movie:
2.2 Mbps x 7200 seconds (i.e. 120 minutes) = 15840 Mb = 15840/8 MB = 1980MB = 1980/1000 GB

= 1.98 GB.

So downloading a movie at the best quality means the CDN serves about 2GB of data to the end-user's Netflix- Roku box.

To arrive on the bandwidth costs, lets go with the figures presented in this PBS article about CDN pricing. Disclaimer: This article is more than a year old, and I have been reading about CDN price wars all along. So the current cost of bandwidth may actually be lower than stated.

From the PBS article the costs for streaming a 2GB stream to a user (assuming volume wholesale pricing):

Single server: $0.26

Akamai : $0.32
P2P: $0.0024

Ok first off, I like the P2P number the most but lets ignore that because P2P may not be able to compete in quality with CDNs (See my paper on this). Even if Netflix uses the most expensive Akamai CDN, they are getting away with just $32 cents per movie instead of the dollar-plus cost in the DVD-mailing model. Even if we assume a few more cents of overhead per movie due to the technology costs, I think Netflix is well in the green with this.

The beauty of Netflix's strategy is that they will be able to gradually wean people from the DVD mailing model to the this online content delivery model because of the convenience of the latter. And this without jeopardizing the DVD mailing model because there is no cannibalization here - its perfect migration with one less DVD mailing customer corresponding to one more streaming customer. Every DVD streamed will add up and lead to a drastic reduction in Netflix's operating costs.

Meanwhile, Roku will probably make some money out of their $99 box.

Last question: And the ISP?
Thats for later. Enjoy your movies.

Update: Netflix bandwidth costs come to about $5cents as of June 2009, according to this article.

If there is a Microsoft Yahoo deal then startups will feel the pinch

2008-05-20T06:02:00.007+02:00

Figure 1: Representative acquisitions of Yahoo, Google, and Microsoft (from this blog)

A Microsoft takeover of all or some of Yahoo will be a good thing for Microsoft in it's battle to unseat Google from the Internet's helm. Perhaps a later Microsoft-Facebook arrangement will finally present a credible challenge to Google. I doubt if the Yahoo board will now agree to anything less than the 72% premium over the original share price Microsoft had offered earlier, so Yahoo shareholders will also come out wealthier from the deal. End users are likely to benefit from a stronger alternative to Google as well.

But one quarter will suffer quietly in the short to medium term: startup companies. Yahoo and Microsoft are some of the most prolific startup acquirers (see Figure 1, from this blog). Yahoo merging with Microsoft removes a big buyer for many startup companies. Moreover, Microsoft will have that much less cash (approximately $44B less based on the first MS offer) to throw at startup acquisitions. With the credit supply tightening up and the economy slowing down, you can be sure of a capital drought ahead for many Internet and software startups.

Crude oil demand: India, China, and the USA

2008-05-16T12:39:00.017+02:00

Figure 1: Crude oil imports of India, China, and the USA

Crude oil prices have never been higher (Brent sweet crude is trading at about $125 a barrel on the NYMEX as of this post). Part of the reason is attributed to the growing demand from emerging economies like India and China that is putting upward pressure on the price of oil. Alan Greenspan writes in his book that the annual world demand for for crude oil has grown by 1.6% since the late 80s while the production has only grown by 0.8% or so annually. The gap has lead investors to bid up crude oil futures in anticipation of the tightening supply, further driving up prices as the buffer between supply and demand has narrowed significantly.

I downloaded crude oil import data from the UN data website for India, China, and the USA and plotted it (Figure 1). Unsurprisingly, the USA imports far more crude oil than India or China. It is more interesting to note is that the growth in US crude oil imports has been of the same order or steeper than that of India and China. Therefore, demand is being driven higher more by the USA than by India or China.

In his book, Alan Greenspan speaks about the "crude oil intensity" of a nation, defined as its crude oil consumption normalized by its GDP. He states that this number is far higher for China and India than it is for the USA because the latter has shifted to a less oil-intensive service economy in the past few decades. From my perspective, I think that the real crude oil intensity of the USA may be much more than Greenspan computes it to be because of USA's large number of imports from China. For example, a plastic toy imported from China counts the crude oil used to manufacture it and transport it to the USA as crude oil used by China.

It seems clear that the biggest lever to reduce crude oil demand lies in the hands of the USA. India and China are emerging economies eager to lift 100s of millions of people out of poverty. As such, they may not have the political capital to cut back on their increasing (but still small) usage of crude oil. On the other hand, even a small percentage cutback in the USA will reduce demand significantly. Lets hope that the USA moves towards more efficient cars, better public transport systems and away from its suburban driving culture in order to keep crude oil within reach of poorer nations of the World.

On dear peer-to-peer

2008-05-14T10:36:00.018+02:00

I will be presenting a paper* at the IEEE IwQoS early next month containing the analysis of a large scale peer-to-peer live video multicast streaming session on the Internet. Think of the P2P video multicast system as the Bittorrent for video streaming (instead of file-sharing). The system was sending a video stream of a baseball match to 10s of thousands of viewers on the Internet using P2P technology.

The presentation and paper are available online. Here are two results from the paper that in my opinion warrant particular notice.

Figure 1 shows the aggregate download and upload bandwidth consumed by the P2P system. Note the scale on the Y axis - Gbps! I am wondering, this is only 1 video stream. What happens to the Internet when 1000s of such streams become available online? Were networks designed for such usage?

An interesting artifact in this figure is that the aggregate download rate of all peers exceeds the aggregate upload rate of all peers. The difference was made up through "bandwidth injecting super-servers" of the content provider. Still, it is absolutely remarkable that the amount of additional bandwidth required is almost constant even as the number of peers increase (Figure 2).

Figure 1: Total bandwidth. Click to enlarge.

In Figure 2, you can see the number of concurrent peers in the P2P system over the time-period of the streamed game (hour 4 to hour 8). Look at the rate of change of peers in (peers joining, peers leaving) the P2P system. Keeping in mind that most of the bandwidth comes from these very peers, it is remarkable that this highly dynamic pool of peers is able to sustain the P2P system. Things get very exciting at the end of the baseball game (Hour 8): Everybody wants to leave. Now that is a big challenge for any P2P system.

Figure 2: Peer Dynamics. Click to Enlarge.

*Joint work with Jatinder Pal Singh (T-Labs) and Aditya Mavlankar, Pierpaolo Baccichet, and Bernd Girod (Stanford University)

Spinning a jetliner...in a loom!

2008-05-10T16:49:00.013+02:00

There was an incredible article in Fortune about the manufacturing process of the Boeing 787. The first plane is slated to fly before the end of the year and Boeing is already reporting that the 787 is the fastest selling jetliner of all times. In fact, Boeing is taking flak from customers for delaying the delivery of the aircraft for want of parts. Apparently, a handful of suppliers just cannot keep up with the demand!

But what fascinated me was that the aircraft's shell will not be made out of aluminum alloys. Instead, it is made of carbon composites. This material is created from carbon fibers that are spun in a way reminiscent of spinning thread (See picture below). Epoxy raisins and subsequent heat treatment creates the carbon composite material for the plane's body. The material is lighter, stronger, and amenable to better aerodynamic design. For example, the entire toilet of the 787 weighs just 170 lb!

Spinning the carbon fibers for the Boeing 787 (courtesy, Fortune). The complete slide-show is available here.

Needless to say, the light aircraft means it is significantly more fuel efficient. At the same time, the greater strength means that the cabin can be pressurized to about 6000ft, making flying more comfortable. The material also allows more comfortable humidity levels in the cabin because unlike metals, carbon does not corrode.

Kudos to the Boeing engineers for designing this marvel!

The recession and US housing prices early in the decade

2008-05-09T06:47:00.007+02:00

Figure 1:Median and Average Sales Prices of New Homes Sold in United States. Up, up, and then recently, down.
Click to Enlarge

I came across an interesting statement in Alan Greenspan's book (1st ed., p 225) explaining the reasons for the relative mildness of the 2001 US recession :

"In the United States, homes had increased in value so much that households, feeling flush, seemed more willing to spend"

It is widely acknowledged now (2008) that housing prices were inflated artificially due to the easy credit available in the first half of the decade. For example, Figure 1 shows the average and median US home prices between 1980 and 2007 (source: US census). Notice the very rapid increase in new home prices between 2000 and 2006. And the most recent slowdown.

Greenspan has touched on the fundamental force responsible for the 2001 recession being so mild - the economy's white knight in shining armor was the US consumer with pockets full from the soaring house prices. All those refinancing dollars kept up the consumer spending. The 2000-01 dot com debacle was perhaps larger than it seemed; but the cushioning effect of housing prices made the pain a lot less, back then.

Does this mean that it is payback time now? Did we end up borrowing from the future in 2001? With housing prices plateauing and actually decreasing in some markets, I very much doubt of housing will bail out the economy this time.

Wanted: New knight in shining armor, preferably resilient to speculative forces.

Well I've read that the weak dollar is good for US exports and also makes foreign imports dearer (e.g. $120+ oil). Perhaps these two factors will buttress the US economy this time. Wait and watch.

Alan Greenspan's book

2008-05-08T11:55:00.019+02:00

The Guru's book (click to visit Amazon site). Picture courtesy Penguin.

I have been reading Alan Greenspan's "The Age of Turbulence: Adventures in a New World" (ISBN 1594201315) over the past couple of days. This book is part autobiography, but given the stature of the man, the book gives an unique insight into how the Global economy has progressed since the War from the perspective of the most well known central banker of all times.

In the book, Alan declares his unflinching faith in Adam Smith's capitalist ideas. Interestingly he admits being deeply influenced by Ayn Rand's (sometimes extreme) beliefs in laissez-faire capitalism. Perhaps the fall of the Soviet system prior to his writing the book stoked the ferocity of Alan's pro-market capitalism beliefs in the book. Whatever the reasons, the fact that market capitalism remains the only time-tested successful economic system makes me comfortable with much of what he says on this topic.

But the book has a lot more to it. One of the interesting features is the interplay of monetary policy and politics in Washington described in the book. It is absolutely remarkable that Alan Greenspan successfully navigated a continuous 18-year term as Fed Chairman with presidents from either side of the political spectrum. Although Alan is a republican, he plainly states his poor opinion of George Bush and extols Bill Clinton"s economic policies. I believe that his ability to work with multiple administrations may also explain his ability to successfully tackle situations like the 87' stock market crash and the post 9/11 economic landscape.

The book explains the basic workings of the Fed in an easy-to-comprehend manner. You learn of technicalities like the Fed Fund rate and the Discount rate, the organizational structure of the Federal Reserve and its relation to the US Treasury, its mandate, etc. While this information is also available elsewhere, Alan brings facts to life by introducing the workings of the Fed through anecdotes and examples from his experience.

I am still on page 200. This is one of those rare books about which I hesitate to comment more before better understanding what the Guru wants us to hear. Stay tuned for more commentary on this superb book.

TATA's acquisition spree. Answer to the "why" question

2008-04-19T20:10:00.016+02:00

I have known Tata since I was a child. I rode Tata buses to school, soaped myself with Tata soaps, stayed in Tata hotels and probably lived in houses supported by Tata steel. More recently both my brother and my wife worked for a Tata company. But the Tata I knew then was different than today's Tata. The global, competitive, aggressive, and ambitious Tata.

India's Tata group has become very well known in Europe recently after taking over UK's Corus and Ford's Land Rover and Jaguar businesses. It is also in talks to buy T-Systems from Deutsche Telekom in Germany. If you add to all this the excitement of the €1800 Nano car designed by Tata, you have got a credible Asian multi-national company in the world's eyes.

Many question the business sense of taking over Landrover and Jaguar given that the premium attached to these brands has a minuscule market in India, where a Hyundai Sonata is considered a luxury salon car! Others question the ability of the Tatas to control unionized European operations.

But I am bullish about their recent acquisitions and their ability to turn these into strategic wins. The foremost reason for my belief is the relative professionalism in the way Tatas conduct their business. For example, top Tata managers of the holding company, Tata Sons, are selected and groomed through the Tata Administrative Service (TAS), admission to which is based on a merit-based, competitive and through screening process. So we can be certain that folks running the acquisition show from the Tata side will be competent and highly trained.

The second reason for my bullish assessment is the value of the technology transfer from Jaguar and Landrover to Tata. About 8 years ago Tata motors rode a huge success in the "Tata SUMO", a rugged diesel-powered 4x4 that looked like a SUV but cost a lot less. The SUMO was an instant hit because it rode well on India's broken roads and because the government of India subsidizes diesel, it was highly cost effective. I still remember seeing caravans of hired SUMOs on highways during weekends and vacations.

Then Toyota came and stole the show with the Toyota "Qualis". This gem had more to offer: Toyota quality, quieter engine, superior interiors, and better fuel efficiency. Although the Qualis was priced slightly higher than the SUMO, it quickly overtook the latter. Tata had lost out because Toyota had superior technology. Lesson learn t for Tata: India had stepped out of its socialist past and now quality and technology mattered to Indian customers.

Jaguar and Landrover will fill this important gap for Tata. Another example where Tata can benefit from better technology is Tata's subcompact "Indica". I have ridden this car and can confidently state that it is noisier and bumpier than its Suzuki counterpart on Indian roads. Jaguar and Landrover technology will perfectly compliment Tata technology. A company that produces some of the cheapest steel in the world (because Tata owns iron and coal mines in India) combined with an established brand and distribution network in India will be unbeatable with the infusion of the latest technical know how from the acquisitions.

Beyond desktop search...can we make user-PC data more valuable?

2008-04-06T21:47:00.007+02:00

I ordered an 8GB flash disk last week (turns out there is a 16GB one around, but I am modest ;-) ). Since I don't have a whole lot of media content to put on this fat-stick, I will instead end up putting all my work over the past few years on it. If I factor in the emails, I should easily fill up 8GB with a couple of years' worth of data. Wow. I remember having a hard time filling a DSDD 5.25 inch 576 kB diskette back in the early 90s.

The low down is that we have *lots* of data. 7MP pictures, podcasts, email archives, documents and web-downloads, not to mention audio/visual media - all this can quickly add up. Fortunately storage has kept up, or perhaps the pace of storage encourages more data generation in the first place? Whatever the truth, we have a situation where we have a whole lot of data sitting in our computers.

There have been many instances of large volumes of data in non-PC computer scenarios. For example, real databases have routinely run into terabytes. The key difference between user-PC data and these databases is the heterogeneity and the lack of structure in the former. User-PC data comes in various formats and is generated by completely different applications. Even when the same application generates the data (e.g. an email mailbox file) the goal has never been to store the data in a way extract global information later.

Desktop search software is the first step in mining information from User-PC data. But search is really a very preliminary tool because it only flags the existence of the information sought via the specified key-words. There is very little cognizance of the bigger picture. Data mining - that power tool which works so beautifully for databases and other highly structured data - does not exist yet for User-PC data.

Isn't it time we started building algorithms beyond just search to help users extract useful information from their gigabytes of data?

ASUS EEPC horror story

2008-03-28T09:10:00.004+01:00

Last week I was at the Tridentcom conference in Innsbruck. There, while attending a talk, I saw first-hand the horror of a presentation gone completely wrong, the discomfort of the audience, and the pain of an embarrassed speaker as his ASUS EEPC failed miserably. The presenters beautiful slides were cut horizontally (about 30% of the lower part of each slide was missing) because the ASUS EEPC could not drive the overhead projector properly. Every other laptop, including old clunky student-budget ones, running Windows , Linux, and Mac OS, were successful in beaming presentations. But not the ASUS EEPC.

There are many objective reviews of laptops on the Internet based on specifications, design, speed, etc. They try to compare products side by side so that potential buyers can choose the product that is best for them. Still, somethings are considered standard and not even mentioned - like the assurance that the power adaptor will charge the laptop's battery, the battery won't burst, the USB ports will work, and the VGA output will work. Seasoned customers are smart enough to get a mix of online and off-line opinions about a product before buying it. But sometimes when products are just released, it is hard to ascertain whether a product will serve its purpose down the road. I hope this EEPC horror story gives folks some additional information before they buy it.

Its all nice to tout the small, cheap ASUS EEPC. But seriously, didn't VGA-out technology mature like 15 years ago? And ASUS cannot even get this right in 2008? My 2 cents - get a second or third hand Pentium 3 laptop or something instead.

Recursion and "Towers of Hanoi"

2008-03-24T07:41:00.031+01:00

Figure 1: rules of the Hanoi puzzle. Click to enlarge

I was thinking of my first algorithms class and remembered the "Towers of Hanoi" problem. This problem is used to introduce the concept of recursion. It can also use the stack data structure (LIFO) quite naturally, making it a universal favorite in assignments. An Internet search yields 1000's of websites discussing the problem. Heres adding yet another discussion to this timeless classic!

Problem Setup
Here is the setup and the rules of the Hanoi puzzle

There are 3 towers labeled 0,1, and 2.
There are N circular discs, each of different radius, that are initially placed around tower 0.
Any circular disc can only be placed above a disc of larger radius, or as the first disc of the tower.

The objective is to move the N discs from tower 0 to tower 2 following the rules mentioned above.

Figure 1 explains the problem graphically and shows the rules of the Hanoi puzzle.

Solution
In order to solve the puzzle recursively, look at Figure 2. The first requirement on moving the largest (red) disc from tower 0 (source tower) to tower 2 (destination tower) is that there should be no discs over this red disc. This means that the other 2 discs (green and yellow) should be moved to tower 1 (auxiliary tower) and the setup should be in state 1. Then, the red disc can be moved to tower 2 (state 2).

Now the problem is reduced to one with just 2 discs (yellow and green). These need to be moved from tower 1 (the source) to tower 3 (the destination) using tower 0 as a temporary go-between (auxiliary tower).

Figure 2: Top level steps. Click to enlarge

But wait, how do we accomplish the transition from state 0 to state 1?
Figure 3 shows the intermediate steps to accomplish this. The aim is to move the green and yellow discs from tower 0 (source) to tower 1 (destination) using tower 2 as the auxiliary tower.

Figure 3: Mini-steps from state 0 to state 1. Click to Enlarge

So the high-level idea is to have a function that recursively places the discs in the appropriate towers. This will be clear in the C++ "Hanoi" function shown below. I have used STL data-structures (vector and stack) in order to simplify the explanation.

#include <stack>
#include <vector>

#include <iostream>

using namespace std;

//Each Tower is a Stack (LIFO)
vector <stack <int>*> towers; //Vector of pointers to stacks

//Printing the tower contents. This is only for eye-candy.

void PrintTowers()
{
stack <int> tempStack;
for (int ii=0;ii<towers.size();ii++) {

 cout << std::endl<< "Tower " << ii <<": ";
 while (!towers[ii]->empty()) {

  tempStack.push(towers[ii]->top());
  towers[ii]->pop();
 }

 while (!tempStack.empty()) {
  towers[ii]->push(tempStack.top());

  cout <<tempStack.top()<<"  "; 
  tempStack.pop();
 }
 cout << " <-- TOP";
}
}

//Move from Tower "from" to Tower "to"
void MoveDisc(int from, int to)
{
towers[to]->push(towers[from]->top());

towers[from]->pop();
}

//Recursive function Hanoi - the most interesting function
//When this function exits, it has moved the lowest
//Tower (as specified by numDiscs) from the sourceTower to the
//destTower using the auxTower as a temporary holder
void Hanoi(int sourceTower, int destTower, int auxTower, int numDiscs)
{

if (numDiscs==0)
 return;
Hanoi(sourceTower, auxTower, destTower, numDiscs-1);

MoveDisc(sourceTower,destTower);
Hanoi(auxTower,destTower,sourceTower,numDiscs-1);
}

int main()
{
int numDiscs;

cout << "Please enter the number of discs: ";

cin >> numDiscs;

//Initialize the Towers by allocating 3 stacks
 for (int ii=0;ii < 3;ii++)

 towers.push_back(new stack<int>);

//Initial condition: all the discs are on Tower 0
 for (int jj=numDiscs-1;jj >=0 ;jj--)

 towers[0]->push(jj);

//Print setup of towers
 cout <<std::endl<<"Initially";

PrintTowers();

//Enter recursive function here
 Hanoi(0, 2, 1, numDiscs);

//Print setup of towers
 cout <<std::endl<<"Finally";
PrintTowers();

//Don't forget to delete the vector of stacks.

 for (int ii=0;ii < 3; ii++)

 delete towers[ii];

cout << std::endl << "Done.";


return 0;
}

Our world needs electricity, lots of it.

2008-03-22T17:33:00.006+01:00

(Click to enlarge)

I visited the UN data website and downloaded data about the total electricity production of countries and matched it with their populations in order to obtain the electricity produced per 1000 inhabitants (for each country). I have plotted a histogram of this in the figure above.

This histogram is disturbing.

The smaller issue is that inhabitants of most countries have lesser electricity than the world average. This means that a few energy-rich countries are producing (and probably consuming) most of the world's electricity. I have marked some of the representative countries on the histogram. There is a table at the end of this post containing the parsed data in case you want to look up your own country and/or use the data (After citing the UN data source, off course).

The bigger issue is that countries with low electricity per capita (bars toward the left of the figure) are striving to move the per capita electricity production higher to improve quality of life. For example, both India and China are lower than the world average. And I have the impression that these countries are really looking to improve their citizen's living conditions. And remember, populations are increasing too (see one of my previous posts), making it necessary to pump up electricity production even faster.

Producing more electricity is a positive development. But the problems associated with increasing production are the issue here. Here are some questions:

Where are we going to get the energy to increase electricity production so rapidly?
What will be the environmental cost of creating so much additional capacity? I hope it is renewable energy. Hope. Hope. Hope.
Or, does this analysis indicate that even in the next decades electricity will be a premium, for-the-well-off, limited quantity luxury given the lack of such a massive energy source?

Tough cookies.

We really need a breakthrough with some new technology here.

Here is the table containing data used in the analysis.
Country, million kWh per 1000 inhabitants (German notation: "," is the decimal point)

Afghanistan	0,019507403
Albania	0,529214445
Algeria	0,219150336
American Samoa	0,936753525
Angola	0,032618392
Anguilla	1,036227154
Antigua and Barbuda	0,325148424
Argentina	0,727408376
Armenia	1,07765584
Aruba	1,457768448
Australia	2,492244294
Austria	2,280999506
Azerbaijan	0,617455344
Bahamas	1,40738335
Bahrain	2,551090802
Bangladesh	0,027811644
Barbados	0,739895798
Belarus	0,819169464
Belgium	1,54779036
Belize	0,174199589
Benin	0,006949106
Bermuda	2,726961075
Bhutan	0,549439336
Bolivia	0,149749265
Bosnia and Herzegovina	0,69957433
Botswana	0,118195712
Brazil	0,49861704
British Virgin Islands	0,454215116
Brunei Darussalam	2,030329213
Bulgaria	1,545853099
Burkina Faso	0,005598074
Burundi	0,004199119
Cambodia	0,013614697
Cameroon	0,049170704
Canada	3,765512578
Cape Verde	0,157851016
Cayman Islands	2,346954443
Central African Republic	0,010259031
Chad	0,002858379
Chile	0,798215316
China	0,386906459
Colombia	0,296546128
Comoros	0,006266434
Congo	0,025762836
Cook Islands	0,57208238
Costa Rica	0,44965969
Croatia	0,849392177
Cuba	0,379683488
Cyprus	1,34517727
Czech Republic	1,708438639
Democratic People's Republic of Korea	0,402276274
Democratic Republic of the Congo	0,043802793
Denmark	2,471503772
Djibouti	0,146728575
Dominica	0,353841391
Dominican Republic	0,582812306
Ecuador	0,273103278
Egypt	0,280151791
El Salvador	0,184723191
Equatorial Guinea	0,026854067
Eritrea	0,036892038
Estonia	1,702729723
Faeroe Islands	1,804792034
Falkland Islands (Malvinas)	3,025210084
Fiji	0,199264292
Finland	3,139151247
French Guiana	0,728790884
French Polynesia	0,442041685
Gabon	0,325406584
Gambia	0,018552543
Georgia	0,977777798
Germany	1,51273341
Ghana	0,065320583
Gibraltar	1,202749141
Greece	1,20759618
Greenland	1,844280122
Grenada	0,304075563
Guadeloupe	0,937493585
Guam	3,274604022
Guatemala	0,162208554
Guinea	0,022771058
Guinea-Bissau	0,01315024
Guyana	0,415837246
Haiti	0,023772922
Honduras	0,212566084
Hungary	0,856203515
Iceland	5,200654647
India	0,126738013
Indonesia	0,116874477
Iran (Islamic Republic of)	0,639435492
Iraq	0,300043035
Ireland	1,518598487
Israel	1,541832479
Jamaica	0,425354403
Japan	2,168335174
Jordan	0,389966498
Kazakhstan	1,231640364
Kenya	0,034242581
Kiribati	0,032607632
Kuwait	4,02037037
Kyrgyzstan	0,710476911
Lao People's Democratic Republic	0,08121598
Latvia	0,940571111
Lebanon	0,601385281
Liberia	0,054622645
Libyan Arab Jamahiriya	0,865970274
Lithuania	1,330189073
Luxembourg	3,631083653
Madagascar	0,012230063
Malawi	0,01376068
Malaysia	0,829451232
Maldives	0,165934635
Mali	0,0098182
Malta	2,28753381
Marshall Islands	0,303244006
Martinique	1,000262695
Mauritania	0,058047217
Mauritius	0,554314346
Mexico	0,490100396
Mongolia	0,322392649
Montserrat	1,776830135
Morocco	0,172225006
Mozambique	0,115377076
Myanmar	0,024954518
Namibia	0,034659007
Nauru	0,989021857
Nepal	0,022551405
Netherlands	1,3307455
Netherlands Antilles	1,126657796
New Caledonia	1,528705938
New Zealand	2,168356638
Nicaragua	0,099770455
Niger	0,007916051
Nigeria	0,041604152
Niue	0,612745098
Oman	1,189848435
Pakistan	0,123101766
Palau	2,583594177
Panama	0,508432302
Papua New Guinea	0,117468448
Paraguay	1,25604174
Peru	0,228699463
Philippines	0,185003073
Poland	0,844522287
Portugal	1,269159686
Puerto Rico	1,370991383
Qatar	3,553189833
Republic of Korea	1,389956686
Romania	0,876196974
Russian Federation	1,619416414
Rwanda	0,004223616
Saint Helena	0,625097672
Saint Kitts and Nevis	0,407016973
Saint Lucia	0,40932771
Saint Pierre and Miquelon	4,254648598
Saint Vincent and the Grenadines	0,29377943
Samoa	0,157741576
Sao Tome and Principe	0,032760677
Saudi Arabia	1,420230761
Senegal	0,048341849
Seychelles	1,110695412
Sierra Leone	0,008771297
Singapore	2,347562131
Slovakia	1,624096551
Slovenia	1,496430224
Solomon Islands	0,02963471
South Africa	0,835171394
Spain	1,888334973
Sri Lanka	0,126093294
Sudan	0,030204543
Suriname	0,859729307
Swaziland	0,122718045
Sweden	3,694381387
Syrian Arab Republic	0,398277093
Tajikistan	0,678298553
Thailand	0,52931205
The former Yugoslav Republic of Macedonia	0,767583489
Timor-Leste	0,042163059
Togo	0,007694068
Tonga	0,080514488
Trinidad and Tobago	1,118059532
Tunisia	0,326482221
Turkey	0,532316671
Turkmenistan	0,811252681
Turks and Caicos Islands	0,163538984
Uganda	0,010743706
Ukraine	1,119794335
United Arab Emirates	3,827701301
United Kingdom	1,358755508
United Republic of Tanzania	0,01426794
United States of America	3,558514712
Uruguay	0,629035396
Uzbekistan	0,440301803
Vanuatu	0,055719101
Viet Nam	0,140517355
Western Sahara	0,131690083
Zambia	0,196892977
Zimbabwe	0,178739129

GENI and Networking Research: Inside out, or outside in?

2008-03-19T14:29:00.007+01:00

The Global Environment for Network Innovations (GENI) idea is to build an experimental facility that researchers can use to experiment with new communications and networking technology, distributed systems, cyber-security aspects, and applications. A couple of weeks ago I attended a talk given by Craig Patridge, chief scientist at BBN technologies. Craig is heading the GENI project office tasked with implementing GENI. He is an engaging speaker and got me thinking about the merits of building GENI.

GENI is a sort of first for the National Science Foundation (NSF) and the network research community. NSF occasionally funds large infrastructure projects like building astronomy telescopes and particle accelerators but this is the first time that a networking infrastructure is being funded. The interesting thing is that the infrastructure comes first and then protocols or services follow. Moreover, the infrastructure will be built based on requirements specified by the research community. Is this the future of network research?

The current Internet is an engineering marvel. It scaling property is absolutely remarkable - 100s of millions of hosts in a federated environment with completely different underlying access technologies just work. Then the creative engineers and innovators come in to shape the Internet APIs (e.g. IP communication stack) into myriad applications - email, VOIP, P2P, Web 2.0, digital libraries, and whatever else they can think of.

The Internet was first built by engineers, and later scientists highlighted several flaws in its design. This has sometimes served as a good feedback loop for refining the Internet over time. For example, security researchers have continually unearthed security holes in the IP socket stack. Scientists, coming from the outside, study the insides of the Internet and contribute to refining the already-built Internet.

But can researchers with limited engineering experience design a new communication infrastructure such as the GENI infrastructure? I very much doubt this. Researchers are seldom successful in building viable commercial technologies. They are very smart but usually focused on one or few problems. It is not at all clear if GENI could deliver the next Internet.

But lets back-pedal a little. Is GENI supposed to be building the next Internet? Answer: perhaps not. But what I find troubling about the GENI (or FIRE - the European counterpart) is that there is that almost the whole OSI stack - network, transport, session and applications - is supposed to come from the research community. I really really wonder who is going to write all this up? Off course there can be a module-based approach to plugging in pre-existing pieces into GENI, but then how is this Internet design revolutionary, and does this justify building GENI in the first place? Why not stick with something more real like PlanetLab?

I don't believe scientists can build another Internet from the inside out. Shouldn't building a new network, from the inside, be left to the engineers?

Bullish about mobile/cellular networks

2008-03-16T13:25:00.003+01:00

(Click to enlarge. Source: UN data)

We all know that cellular mobile networks are rapidly expanding all over the world - in fact the rate of adoption of cellular mobile technology has been faster than that of the Internet. I am bullish about this technology's role in human development. The bar-graph says it all. Amen!

The Internet gets a heavy tail

2008-03-08T11:23:00.017+01:00

(Click to enlarge)

The above plot shows the growth of the Internet in the World, again from UN data. There are 3 histograms (2004, 2001, 1998) of the number of countries (Y axis) vs. the size of the Internet population in the countries (X axis). I have left out the countries with less than 1m users because they are the vast majority in the data - partly due to low Internet adoption in backward areas and partly due to small populations.

The key points seen immediately

The number of Internet users is (surprise surprise) growing rapidly over the years.
USA leads every other country, but since the total US population is only about 300m (see my previous post), US Internet users' growth is saturating.
On the other hand, China is growing very quickly. I suppose by 2008 (present) the number of users there must have got close to the US number (if not ahead).

But lets look at a more holistic trend here. This is the "long fat tail" that shows up in the 2004 data - the bars concentrated on the left hand side have started spreading out to the right in the 2004 histogram. This means that the Internet is increasingly becoming multinational with a significant representation of people from many countries, bringing up these questions

Where is the Internet content suitable for people from different cultures and not just western cultures?
Are we doing enough to localize Internet applications for these large multicultural groups joining the Internet?

Finally, for those interested, here is a table showing the number of Internet users, as reported by the most recent data from UN Statistics (2004). I have ranked the countries based on the number of Internet users in the rightmost columns of the table. Again, if you look at the top 10 countries in this table, only 2 have a high number of English native speakers. In addition, 5 out of the top 10 countries are not "Northern hemisphere western countries".

Please use this data only after citing the UN source.

Country	#Internet users	Rank (2004)
United States	185000000	1
China	94000000	2
Japan	64160000	3
United Kingdom	37600000	4
Germany	35200000	5
India	35000000	6
Korea, Republic of	31580000	7
Italy	28870000	8
France	25000000	9
Brazil	22000000	10
Canada	20000000	11
Russian Federation	16000000	12
Indonesia	14508000	13
Spain	14332800	14
Mexico	14036500	15
Australia	13000000	16
Turkey	10220000	17
Netherlands	10000000	18
Malaysia	9879000	19
Poland	9000000	20
Thailand	6972000	21
Sweden	6800000	22
Argentina	6153600	23
Viet Nam	5870000	24
Iran (Islamic Republic of)	5500000	25
Czech Republic	5100000	26
Romania	4500000	27
Philippines	4400000	28
Chile	4300000	29
Belgium	4200000	30
Colombia	4050240	31
Austria	3900000	32
Egypt	3900000	33
Ukraine	3750000	34
South Africa	3566000	35
Morocco	3500000	36
Switzerland	3500000	37
China, Hong Kong Special Administrative Region	3479700	38
Finland	3286000	39
Peru	3220000	40
Israel	3200000	41
Portugal	2951000	42
Denmark	2725000	43
Hungary	2700000	44
Belarus	2461090	45
Singapore	2421780	46
Venezuela	2312680	47
Slovakia	2276060	48
New Zealand	2110000	49
Pakistan	2000000	50
Greece	1955000	51
Norway	1792000	52
Nigeria	1769660	53
Saudi Arabia	1586000	54
Serbia and Montenegro	1517020	55
Kenya	1500000	56
United Arab Emirates	1384840	57
Croatia	1303000	58
Bulgaria	1234000	59
Ireland	1198000	60
Sudan	1140000	61
Jamaica	1067000	62
Costa Rica	1000000	63
Lithuania	968000	64
Slovenia	950000	65
Uzbekistan	880000	66
Puerto Rico	862000	67
Tunisia	835000	68
Zimbabwe	820000	69
Latvia	810000	70
Dominican Republic	800000	71
Syrian Arab Republic	800000	72
Guatemala	756000	73
Uruguay	680000	74
Estonia	670000	75
Ecuador	624579	76
Jordan	600000	77
Kuwait	600000	78
Lebanon	600000	79
El Salvador	587475	80
Algeria	500000	81
Haiti	500000	82
Senegal	482000	83
Azerbaijan	408000	84
Republic of Moldova	406000	85
Kazakhstan	400000	86
Ghana	368000	87
Bolivia	350000	88
United Republic of Tanzania	333000	89
Malta	301000	90
Bangladesh	300000	91
Panama	300000	92
Cyprus	298000	93
Sri Lanka	280000	94
Luxembourg	270810	95
Kyrgyzstan	263000	96
Oman	245000	97
Cote d'Ivoire	240000	98
Zambia	231000	99
Iceland	225610	100
Bosnia and Herzegovina	225000	101
Honduras	222273	102
Togo	221000	103
Libyan Arab Jamahiriya	205000	104
Mongolia	200000	105
Reunion	200000	106
Uganda	200000	107
Mauritius	180000	108
Yemen	180000	109
Georgia	175600	110
Angola	172000	111
Papua New Guinea	170000	112
Cameroon	167000	113
Qatar	165000	114
Occupied Palestinian Territory	160000	115
Trinidad and Tobago	160000	116
The former Yugoslav Republic of Macedonia	159000	117
Bahrain	152721	118
Armenia	150000	119
Barbados	150000	120
China, Macao Special Administrative Region	150000	121
Cuba	150000	122
Paraguay	150000	123
Guyana	145000	124
Mozambique	138000	125
Nicaragua	125000	126
Nepal	120000	127
Ethiopia	113000	128
Martinique	107000	129
Benin	100000	130
Bahamas	93000	131
Madagascar	90000	132
Guadeloupe	79000	133
Guam	79000	134
Namibia	75000	135
New Caledonia	70000	136
Myanmar	63688	137
Fiji	61000	138
Botswana	60000	139
Brunei Darussalam	56000	140
Saint Lucia	55000	141
Burkina Faso	53200	142
Democratic Republic of the Congo	50000	143
Eritrea	50000	144
Mali	50000	145
Gambia	49000	146
Malawi	46140	147
Guinea	46000	148
French Polynesia	45000	149
Lesotho	43000	150
Cambodia	41000	151
Gabon	40000	152
Bermuda	39000	153
French Guiana	38000	154
Greenland	38000	155
Rwanda	38000	156
Congo	36000	157
Iraq	36000	158
Swaziland	36000	159
Turkmenistan	36000	160
Belize	35000	161
Chad	35000	162
Faeroe Islands	32000	163
Suriname	30000	164
United States Virgin Islands	30000	165
Guinea-Bissau	26000	166
Burundi	25000	167
Cape Verde	25000	168
Aruba	24000	169
Niger	24000	170
Liechtenstein	22000	171
Lao People's Democratic Republic	20900	172
Dominica	20500	173
Antigua and Barbuda	20000	174
Bhutan	20000	175
Sao Tome and Principe	20000	176
Seychelles	20000	177
Grenada	19000	178
Maldives	19000	179
San Marino	15000	180
Somalia	15000	181
Mauritania	14000	182
Micronesia, Federated States of	12000	183
Andorra	11000	184
Saint Kitts and Nevis	10000	185
Sierra Leone	10000	186
Central African Republic	9000	187
Djibouti	9000	188
Comoros	8000	189
Saint Vincent and the Grenadines	8000	190
Vanuatu	7500	191
Gibraltar	6295	192
Samoa	6000	193
Equatorial Guinea	5000	194
Tajikistan	5000	195
Palau	4000	196
Solomon Islands	3000	197
Tonga	3000	198
Tuvalu	3000	199
Kiribati	2000	200
Marshall Islands	2000	201
Netherlands Antilles	2000	202
Cayman Islands	1300	203
Liberia	1000	204
Nauru	300	205
Korea, Democratic People's Republic of	0	206