Mukul Kumar's Blog

Great startup advice from Ben Horowitz

2012-01-13T11:42:00.000+05:30

Ben Horowitz gave some great advice today at the 'Geeks vs BizDev in startups' event at Mountain View. I followed it on twitter and have curated the best quotes. Here you go, read them slow :) .

"At a startup you only feel two emotions: fear and euphoria. @bhorowitz #geeksvsuits

"If takes longer to argue then build it, just build it" #geeksvsuits

If you don't know what you want it is very unlikely you will get it - @bhorowitz. #geeksvsuits

"If the idea is a big enough breakthrough, it will attract world-class people" @bhorowitz #geeksvsuits

"Monetization is a byproduct of creating something that has value for a lot of people" -@bhorowitz #geeksvsuits

"The one thing you can't learn in business school is management" @bhorowitz #geeksvsuits

"Trust and communication are equivalent." @bhorowitz #geeksvsuits "If your commutations are breaking start by building trust."

Build a diverse team that argues, trusts each other, and share the same language | Manish Patel from @HighlandCapital #geeksvsuits

Entrepreneur: "Our company is going to stay 15 people forever." @bhorowitz: "Oh, so you're planning to fail." #geeksvsuits

The culture of a winning startup: You win or you die. -@bhorowitz #geeksvsuits

"I'd rather be right than consistent" @bhorowitz #geeksvsuits

"My mother always told me if you ask for a million dollars, be sure to wear a suit." --Ben Horowitz #GeeksVSuits

"You've got to think about why your 20th employee would join your company." @bhorowitz #geeksvsuits

Name me a successful company that has outsourced engineering. - None, that was a trick question! @bhorowitz #geeksvsuits

Non-technical founders fail if they don't realize the nature of the tech business and innovate continuously @bhorowitz #geeksvsuits

The best technology doesn't always win but the technology that wins is always the best. @bhorowitz #geeksvsuits

"I'd never do a startup after Amazon.I visited my 3rd grade class: it was nice but I'm not going back" @bhorowitz quoting Bezos #geeksvsuits

"Part of figuring out the product/market fit is selling your product. Especially in the enterprise market." @bhorowitz #geeksvsuits

"The idea of outsourcing engineering was conceived by someone who just doesn't understand." ~ @bhorowitz #geeksvsuits

"The main mistake technical founders make is thinking Sales and Marketing is easy and unnecessary. They're wrong." @bhorowitz #geeksvsuits

"the next Facebook or Google is going to start off looking like a toy" - Elad Gil #geeksvsuits

Mistake of 90s was thinking product doesn't matter, mistake of 2000s is product is all that matters -@bhorowitz #geeksvsuits

You want people who are ambitious for the company, not for themselves. ~ @bhorowitz #geeksvsuits

"I don't think I would do a PhD now because it's four years of not doing anything interesting." - @eladgil #startupeducation #geeksvsuits

"Early Twitter went for years without needing anybody who wasn't an engineer." ~ @bhorowitz #geeksvsuits

"FB has 2000 salespeople." @bhorowitz #geeksvsuits

More depth with smaller ISO in DSLR photography

2011-04-17T23:26:00.001+05:30

I did some comparative experimentation of ISO 100 to ISO 200 in daylight photography. I think I can see clear difference between the pictures. Much of this may already be documented, but this is what I found first hand experimentation - at ISO 100 I can see - a) more depth, b) more clarity - a much better picture. ISO 200 gave me - a) a flatter picture, therefore lesser depth, b) less clarity.

See for yourself.

ISO 100

ISO 200

Google getting hypersocial

2011-03-15T17:18:00.001+05:30

While doing Google search, I just noticed this new suggestion today, Google knows about my Twitter id and wants to suggest me what my friends are talking about. See the box which says "Want to see which results your friends ..." I don't think I have told Google about my Twitter id, but they somehow know it. Looks pretty cool though.

When I click on the button they add my Twitter id to my profile.

I don't see any changes in results.

Was it that Google was testing some new feature?

PubMatic Pune Mixer on 26th November

2010-11-23T19:05:00.000+05:30

On the 4^th anniversary of PubMatic, we are holding a networking event at our Pune office in Baner. Please find the details of the event on the following URL - http://pubmaticmixer.eventbrite.com/ . Register for the event here.

It is the second event of this kind that PubMatic has organized. The previous mixer was a great success (read about it here) and we would be delighted to celebrate this occasion with you.

At the PubMatic Mixer Event we will have an informal discussion on industry trends in the online domain followed by a networking dinner.

Suvir Sujan, our investor and Co-founder at Nexus Venture Partners, will also be at this event. Suvir has been Co-founder and Co-CEO of Baazee.com, which merged with eBay in 2004 to form eBay India.

Register for the event here.

Does IT outsourcing industry fear cloud computing?

2010-11-11T15:25:00.001+05:30

I read a comment this morning from Vivek Wadhwa, regarding adoption of cloud computing. Just thought I will post my comments on that.

Here are my thoughts; I think there are 3 reasons:

There is a lot of value (money) in consulting/designing a technical stack, like which servers to buy, which networking equipment etc. All that goes away with a cloud. Cloud gives you all that on demand and on a click
There are a lot of supporting technologies such as CDN, load-balancing, DNS etc., that requires a consultation/designing/deals, which all goes away (see CloudFront for example, or EBS load-balancer), so less work for the consultation guy
There are peripheral services such as infrastructure management, uptime management etc. That is also taken away by the cloud service providers to a certain degree.

Cloud Computing makes the entire process of acquiring IT services, very easy, “at the click of a button”. And that can take away a few jobs.

Concepts 2010 at PICT Pune

2010-04-11T23:31:00.002+05:30

On 27^th March I went to Pune Institute of Computer Technology (PICT) for judging Concepts'2010. I was accompanied by Anand, Ananda, Dinesh and Purnima. "Concepts" is an annual national-level technical event, organized under the joint aegis of PICT and IEEE (PICT Students Chapter, Bombay Section). This event has been organized for the past nineteen years, and has carved a niche for itself as being one of the most rewarded National Technical Events. At the heart of the event is a national level Technical Project competition, which gives students a platform to final year engineering students pursuing Computer Science, Electronics & Telecommunications, and Information Technology and related Engineering fields.

The hospitality provided by the PICT students and staff was superb and beyond expectation. The management of the entire event was flawlessly executed by the students, and it felt truly world class.

[Following is a picture of Ananda, myself and Anand Das - excited after seeing so much technology.]

I reviewed four projects that were done very diligently by the Engineering students. The students had deep understanding of the concepts and were very clear on their fundamentals. They were deeply knowledgeable of the subject matter and would smoothly answer my questions. The projects that I reviewed were XdroidFS: an extension to YAFFS2 (root file-system on Android), Virtual Tape Library, Hierarchical Storage Management (across hard-disk and SSD) and Fusion FS. The students were really innovative in their approach, had designed the solution very well, written enormous amount of code and had fully working prototypes. Pretty amazing work!

I can very well imagine XdroidFS becoming a part of the Android file-system and being used for securing files using xattr. The students doing the Hierarchical Storage Management designed their page tables very well and used HSM techniques to automatically move data between HDD and SSD for fast access and cheaper cost.

[Following is a picture of Anand and Ananda.]

It was great to feel technology again, and to see the passion in the students to solve problems and the drive for innovation.

Great job guys! Absolutely Awesome!

The Return of the Salary Hunters

2010-03-21T22:47:00.002+05:30

The recession showed us some unexpected hiring times. All hiring rules changed, salary negotiations were non-existent, offer-to-join conversion rates were high, 'great work' was more important than 'great salary'. The end of recession brought a sigh of relief to all of us. Companies are ramping up hiring efforts, campuses are seeing heavy hiring activity and packages have gone up and India Inc. has declared the much awaited salary raises . All this is great.

This also brings back the challenges of hiring. All of the following, which had become history are now back, with a "vengeance":

RETURN OF THE PING-PONG SALARY HUNTERS
Taking an offer and negotiating salary or role is a good thing. But taking an offer letter and showing it to another company to get a better deal, is kind of being overly opportunistic. Now, taking another offer letter and showing it back to the last company that gave the original offer letter (ping-ponging) is plain "evil". It is not just evil, it is a bad idea. For a five-ten % salary hike the person is potentially leaving a bad impression on his future employer.

Remember that it is much more important to work for a great company, for a company that has a solid roadmap, it is much more important to work with a great team, where you can learn for the next X number of years you spend in that company. It is much more important to work for a company that values your work, and rewards not just with a great salary but with accolades and encouragement that go much longer and deeper than fifty-hundred thousand rupees per year.

RETURN OF THE MULTIPLE OFFER HOLDER
Almost all offers recently are met with counter offers, not just one but many. I think it is not a bad idea to negotiate salary with other offers with counter offers, though I think it is somewhat naughty to do that. The priorities should be clear when someone wants to work for a company, whether it is a small salary difference that can pull the person to a different company, or is the vision and the roadmap and the team is more important.

Salary is definitely important; after-all EMIs need to be paid each month! But beyond that, there are many more things that are much more important. One can never compare companies purely by the salary that they offer.

It is always a better idea to work for a great company with a great team and a great vision. The most important, like I said, are the team, the quality of work, the work environment, the company's vision, its market leadership and the learning that it provides.

Think deep and thing wide before you take that plunge.

Real Engineers don't give ETAs

2010-01-24T23:32:00.001+05:30

Recently I had an interesting observation, followed up by a very interesting discussion with a few Engineers (on Skype obviously).

The initial observation is that - "Real Engineers don't give ETAs." ETA meaning estimated time of arrival, often used in project planning to describe the date on which a software release will happen.

Here's why Real Engineers don't give ETAs ---

Real Engineers just get it done.
Real Engineers think "Why waste time in calculating an ETA, when you can code in the same time?"
Real Engineers JUST write code ...
"Real Engineers aren't bothered about the time-line but the quality of their code ..."
"ETAs are at times a good way for an engineer to communicate with a non-engineer :)"
"(asking someone else to give the ETA) It allows the engineer to completely focus on solving the problem at hand without much distraction."

Real Engineers actually came up with yet-another-definition of ETA - "Estimated Time of Amnesia". That's the time when everybody (other than the Engineer) starts forgetting "what was actually promised" ... I would love to know your thoughts on this.

Firefox 3.6 has some interesting featureset

2009-11-02T22:05:00.005+05:30

Firefox 3.6 Beta 1 have some really interesting set of new features. Read this page for the full feature list, I have reproduced an excerpt that I found interesting:

DOM

Drag and drop now supports files: The DataTransfer object provided to drag listeners now includes a list of files that were dragged.

Detecting device orientation: Content can now detect the orientation of the device if it has a supported accelerometer, using the MozOrientation event; see window.onmozorientation for details. Firefox 3.6 supports the accelerometer in Mac laptops.

For XUL and add-on developers

If you're an extension developer, you should start by reading Updating extensions for Firefox 3.6, which offers a helpful overview of what changes may affect your extension. Plug-in developers should read Updating plug-ins for Firefox 3.6.

New features

Detecting device orientation

Content can now detect the orientation of the device if it has a supported accelerometer, using the MozOrientation event; see window.onmozorientation for details. Firefox 3.6 supports the accelerometer in Mac laptops.

Monitoring HTTP activity

You can now monitor HTTP transactions to observe requests and responses in real time.

Web Scalability & Performance: Real Life Lessons

2009-09-06T22:37:00.003+05:30

Following is a presentation that I made at TechWeekend in Pune on 5th September. About sixty hard-core technical geeks were present at the sessions. Following is the presentation that I made. Feel free to share.Web Scalability & Performance
You can reach me on Twitter @mukulneetika .

8 cores and above - Is the race worth it?

2009-04-25T22:24:00.001+05:30

AMD has announced plans to beat Intel to 12 cores, releasing both 8 and 12 core processors, codenamed Magny-Cours, in Q1 2010. It has also announced that it will in 2011 roll out its 32 nm Bulldozer core, which will feature up to 16 cores, running on the new Sandtiger architecture. In short -- AMD plans to beat Intel in the core race.

You may note that Engineers at Sandia National Laboratories did simulation of 8, 16 and 32 cores, and have opined that performance of multi-core machines would level off or even decline beyond 8 cores,due to limited memory bandwidth.

Read more here.

At the heart of the trouble is the so-called memory wall—the growing disparity between how fast a CPU can operate on data and how fast it can get the data it needs. Although the number of cores per processor is increasing, the number of connections from the chip to the rest of the computer is not. So keeping all the cores fed with data is a problem. In informatics applications, the problem is worse, explains Richard C. Murphy, a senior member of the technical staff at Sandia, because there is no physical relationship between what a processor may be working on and where the next set of data it needs may reside. Instead of being in the cache of the core next door, the data may be on a DRAM chip in a rack 20 meters away and need to leave the chip, pass through one or more routers and optical fibers, and find its way onto the processor.

I would really love to run the MT tests that would show the performance at 6 and 8 core. Searching ...

Go for that “Impulse Purchase”

2009-04-12T23:27:00.003+05:30

‘Planned purchase’ is something I have been doing for past several years. It is generally believed that as you “mature” you do more planned purchases versus impulsive purchases. Impulsive purchase has been linked with the “impulsive youth”. Sometimes is has also been associated with immaturity. While, planned purchase is something I have generally been very proud of, I have lately realized the downside of this approach, and I am beginning to think if it is a bad thing.

Planned purchase is invariable associated with – a) massive market research, b) massive product research, c) understanding your own requirements, d) deciding the exact price point for best price performance ratio, e) sometime getting into the microscopic details of the product, for differentiation in a commoditized product, technology or service. All this might take anywhere from a few days to a few months – eventually leading to a “better purchase”.

When you buy something, one of the important things that must happen after the purchase is – “you must derive great sense of satisfaction from your purchase”. I have come up with an approximate relationship between satisfaction and the research you do on the product, the relationship is:

So, I am postulating that the satisfaction is inversely proportional to the amount of research you do on the product. The more you do research, the less satisfaction you get after doing the purchase.

If you are buying an electronic gadget, many times, there are several less documented small features that come in with a gadget – like a quick-access button in a smartphone, a hidden pocket in a backpack or the cool 3.2MP webcam in a laptop. These features almost appear like “serendipity” when you accidentally find it, you feel pleasantly surprised in your new gadget. All that surprise is gone when you have done 4 months of study on all the smartphones available in the market. I feel that this is a big bummer.

An extension of the above rule is that satisfaction is directly proportional to the ‘Impulsiveness of Purchase’. I think there are ample examples of my friends who have purchased stuff with little research, and have been immensely satisfied with their purchase.

What do you think?

A caveat to these rules will be – don’t apply this on large sized purchases. Like don’t apply this when buying a house or a car ?

So, guess what, as of today, I am shifting my slider from the ‘planned purchase’ to the ‘impulsive purchase’ side.

Analysis: Which URL Shortening Service Should You Use?

2009-04-07T21:40:00.003+05:30

Danny Sullivan has a written a great writeup on comparing different URL shortners. The article is really interesting, you should read it here.

Danny presented this table in the article, which is the key document in this article. Check it out.

I also did a quick search of the popularity of each of the URL shortening service. Here are the results:

tinyurl.com	23,000,000
bit.ly	3,840,000
is.gd	3,160,000
tr.im	480,000
cli.gs	144,000
snurl	109,000
kl.am	8,810
short.ie	6,060

Mikz - How different is it from httpd4mobile and mymobilesite?

2009-03-05T00:12:00.004+05:30

TechCrunch reports this story today:

Conveneer, a Swedish mobile startup with offices in Lund, Sweden and Palo Alto, California, closed a $4.5 million venture round, led by the Swedish foundation Industrifonden. Broken Arrow Venture Capital also participated. The company previously raised seed money from the founders and Teknoseed. Conveneer is building a mobile platform called Mikz, which will be able to assign a URL to your mobile phone, making the content on your phone accessible on the Web. In essence, it turns each mobile phone into a Web server. Once your phone has a URL like http://joe.mikz.me, other Web applications and services can ingest the data that is locked in your phone, and also your phone can take advantage of common Web APIs. Mikz can pull information off your phone such as your contacts, GPS coordinates, photos, music, ringtones, and other files. It creates a Web interface for your phone.

I haven't gone too deep into how they do it, but I see that httpd4mobile also does something similar. Httpd4mobile is an HTTP mobile server for Java J2ME mobile devices that enables you to access and control various features such as Camera Picture function, Audio Record function, Contact List, SMS Send, File Download, File Upload, etc.

See also mymobilesite that works on Nokia S60 devices, that allows users to create, share and access contacts, calendar appointments, SMS text messages, emails, phone logs, share pictures, etc.

What do you think?

Microsoft Planning Ad-Supported Model For Office 14?

2009-03-04T09:53:00.001+05:30

Money.cnn.com reported:

An ad-supported Microsoft Office 14?
That's what Microsoft Business Division Chief Stephen Elop said was coming at a presentation to analysts at the Morgan Stanley Technology conference today.

Here are my questions :-)

Is this the online-version of office? Would Microsoft do contextual analysis of the documents to put the ads on Office 14? If yes - then the content is no longer private; Microsoft would know all the content (maybe a software pirate would care less). If not - how would they put ads, how would they target a user?

Looking at the top 20 countries that have highest piracy rates, I can't think if there can be relevant ads for those geographies. See the list yourself here.

Interesting.

Contribute to a Map-Reduce job by simply pointing your browser to a URL

2009-03-04T09:35:00.002+05:30

Igvita.com has published an interesting way of running Mapreduce jobs:

After several iterations, false starts, and great conversations with Michael Nielsen, a flash of the obvious came: HTTP + Javascript!

What if you could contribute to a computational (Map-Reduce) job by simply pointing your browser to a URL? Surely your social network wouldn't mind opening a background tab to help you crunch a dataset or
two!

Instead of focusing on high-throughput proprietary protocols and high-efficiency data planes to distribute and deliver the data, we could use battle tested solutions: HTTP and your favorite browser. It just so happens that there are more Javascript processors around the world (every browser can run it) than for any other language out there - a perfect data processing platform.

Some really interesting comments in there too. Read full text here.

10 Interesting Articles for the Weekend Read

2009-02-28T16:20:00.001+05:30

I found the following articles pretty interesting, check them out:

Startups in 13 Sentences - by Paul Graham

How FriendFeed uses MySQL to store schema-less data

Giz Explains: Why Lenses Are the Real Key to Stunning Photos

Cellular providers want Nokia to drop Skype from cell phones - Two cell service providers in the UK are supposedly up in arms over Nokia's inclusion of Skype software on its N97 handset, and are threatening not to carry the device unless the software is ditched. This stance is not only annoying to consumers who are beginning to like VoIP, but it could also even hurt the carriers' business in the long run.

The Big List Of Search Engines & Their Employees On Twitters

How to Buy Domain Names Like a Pro: 10 Tips from the Founder of PhoneTag.com

10 Free CAPTCHA scripts and services for websites

OCRTerminal - free online Optical Character Recognition service that allows you to convert scanned images and PDF's into editable and text searchable documents.

10 Papers Every Programmer Should Read (At Least Twice)

15 online photo editors compared - compares Flauntr, Fotoflexer, Lunapic, Phixr, Phoenix, Photoshop.com, Picnik free, Picnik premium,
Picture2Life, Pixenate, Pixer.us, Pixlr, Snipshot, Snipshot Pro and Splashup.

Y Combinator’s FathomDB looks pretty neat

2009-02-28T14:01:00.002+05:30

FathomDB provides relational databases under the utility/service model. They say that they automate the low-level DBA tasks (backup/monitoring); and also provide performance analysis tools that facilitate the high level DBA tasks.

Final pricing for the service is still being determined, but the company plans to charge a small (~10-20%) markup over standard EC2 prices.

I think this is pretty neat. It's a great trend. I hope they don't have any 'center point of failure', I hope their monitoring resources and backup locations are widely distributed. I think DB backup brings in a lot of value.

This screenshot at Techcrunch looks pretty interesting:

Carol Bartz on 'getting the house in order'

2009-02-27T13:07:00.002+05:30

Carol Bartz, CEO Yahoo! writes on her blog:

Today I’m rolling out a new management structure that I believe will make Yahoo! a lot faster on its feet. For us working at Yahoo!, it means everything gets simpler. We’ll be able to make speedier decisions, the notorious silos are gone, and we have a renewed focus on the customer. For you using Yahoo! every day, it will better enable us to deliver products that make you say, “Wow.”

Impressive.

The reorg at Yahoo! has been discussed in the following articles:

Microsoft to use Machine Learning software to put servers to sleep when not in use

2009-02-26T14:49:00.002+05:30

SAI reports:

Microsoft is working on new tech codenamed "Marlowe" to build data centers using low-power servers that can intelligently "sleep" and "wake up" -- just like a portable computer.

NY Times: The company has applied sophisticated machine learning software to the Atom-based servers and tracked how they handle search requests on Microsoft Live over the course of a day.

When the software senses a lull in action, it can place large numbers of servers into sleep or hibernate modes so that they consume just 2 to 4 watts instead of the usual 28 to 37 watts. Then, in an ideal set-up, the software can anticipate when more active periods will resume and begin waking up the servers ahead of the incoming search requests. It usually takes the servers about 5 to 45 seconds to jump back into action.

I think this is a great idea, if you can predict the load on the servers, and if you datacenter is already not optimized. With hundreds of cron-jobs running at asynchronous intervals, DB and file systems deciding on their own schedule of page-flushes, and with server load coming from around the globe - it might be a challenge to accurately predict the load. But, that's what software is for.

Watch this space.

'AWS Public Data sets' has full Wikipedia available in TSV format

2009-02-26T14:31:00.001+05:30

'Amazon Web Services Blog' reports that the AWS public data sets has the Wikipedia Extraction (WEX), which is a processed, machine-readable dump of the English-language section of the Wikipedia. At nearly 67 GB, this is a handly and formidable data set. The data is provided is the TSV format as exported by PostgreSQL.

There are a number of other data sets also available, read more here.

They also describe how easily you an use these data sets:

Instantiating these data sets is basically trivial. You create a new EBS volume of the appropriate size, basing it on the snapshot id of the data. Next, you attach the volume to a running EC2 instance in the same availability zone. Finally, you create a mount point and mount the EBS volume on the instance.

Awesome.

Google Apps Status Dashboard looks pretty good

2009-02-26T14:20:00.002+05:30

Google has finally come up with a status dashboard. I had previously reported on other Web Services that do similar status reporting, such as http://status.aws.amazon.com/ , http://status.mosso.com/ and http://heartbeat.skype.com/.

Looks great - as long as all the check boxes stay ticked.

Nokia to enter laptop market?

2009-02-26T11:45:00.001+05:30

jkOnTheRun reports this story -

Reuters is reporting that Finnish handset giant Nokia has admitted they are considering entering the laptop market. In an interview in Finland CEO Pekka Kallasvuo was asked if Nokia plans to make laptops and had this response:
“We are looking very actively also at this opportunity… We don’t have to look even for five years from now to see that what we know as a mobile phone and what we know as a PC are in many ways converging.”

Nokia also reported earlier that they would shrink production and R&D as sales tank, and rapidly shrinking market-share in the smartphone market. Netbook is a growing market, no wonder Nokia wants to get a portion of the pie.

CDNetworks Acquires Panther Express To Speed Expansion In The U.S.

2009-02-25T21:28:00.001+05:30

Dan Rayburn does a detailed writeup on this story here:

This morning, CDNetworks announced that it has acquired Panther Express. Headquartered in NYC, privately held Panther Express has been in the content delivery business since 2005 offering HTTP based delivery services in the U.S and Europe.

Panther's footprint gives CDNetworks quicker access into North America and Europe and allows them to ramp sales much faster. What Panther Express lacks is the reach into Asia, the ability to support streaming media services, including live delivery and access to a large sales and marketing force. CDNetworks has the footprint in Asia, supports streaming of Flash, Silverlight and Windows Media live or on-demand and is an organization of over 400 employees after the inclusion of Panther's team.

PubMatic Launches Open, Real-time Monetization API for Ad Networks and Publishers

2009-02-25T19:54:00.002+05:30

PubMatic's API Allows Ad Networks to Instantly Access Premium Publisher Inventory Resulting in Increased Revenue for Participating Publishers

PubMatic (www.pubmatic.com), an ad revenue optimization company that works with over 5,500 online publishers, announced today the official launch of it's highly anticipated Application Programming Interface (API). The API will increase revenue for both ad networks and PubMatic's publisher clients by allowing an instant and transparent connection between them for single or multiple ad campaigns on demand. The official API launch comes after a successful closed beta period with multiple ad network partners.

Benefits for Ad Networks:

Increased Reach: Access to over 125 million unique users and thousands of websites.
Improved Targeting: Ad networks can leverage expanded targeting options including geography, frequency, user re-targeting, ad tag type, and much more.
Campaign Control: Ad networks have total control over pricing and timing for campaigns

Benefits for Publishers:

Increased Monetization: Allowing more ad networks to instantly access publisher inventory increases publisher ad rates and sell through rates.
Increased Visibility: Dozens of new ad networks can get instant and transparent access to publisher inventory.
Better User Experience: More premium campaigns and better targeting result in users seeing higher quality and more relevant advertising.
Zero Integration: PubMatic publishers have instant access to premium campaigns and ad networks via the API, with zero integration effort.

"Efficiency is absolutely key to improving monetization in the online advertising ecosystem," said PubMatic CEO, Rajeev Goel. "Our API was designed to seamlessly connect publishers and ad networks so they can have total control of their campaigns, which results in an easy, streamlined process that is financially beneficial to both parties. This is the future of online advertising."

Google Announces Pricing for App Engine

2009-02-25T18:33:00.001+05:30

RWW reports:

Here is the new pricing scheme according to Google's blog post:
$0.10 per CPU core hour. This covers the actual CPU time an application uses to process a given request, as well as that for any Datastore usage.
$0.10 per GB bandwidth incoming, $0.12 per GB bandwidth outgoing. This covers traffic directly to/from users, traffic between the app and any external servers accessed using the URLFetch API, and data sent via the Email API.
$0.15 per GB of data stored by the application per month.
$0.0001 per email recipient for emails sent by the application
In general, Google's prices seem to be slightly cheaper and less complicated than Amazon's pricing schemes for using its EC2 and S3 service. It should be noted, however, that Amazon offers a far larger feature set than App Engine. App Engine only supports the Python programming language, while EC2 gives you access to a complete, remotely hosted, on-demand operating system.

Also see previous post on Comparing Clouds: Amazon EC2, Google, AppNexus, and GoGrid.

Safari 4 benchmarked: 42x faster than IE 7, 3.5x faster than Firefox 3

2009-02-25T18:26:00.001+05:30

Holy Jesus!

CNET UK reports ...

Proving itself a staggering 42 times faster at rendering JavaScript than IE 7, our benchmarks confirm Apple's Safari 4 browser, released in beta today, is the fastest browser on the planet. In fact, it beat Google's Chrome, Firefox 3, Opera 9.6 and even Mozilla's developmental Minefield browser.

20 million Chinese sites are served by QZHTTP not Apache

2009-02-25T13:40:00.002+05:30

The following report talks about 20 million Chinese sites are served by QZHTTP not Apache.

Also, Qzone blogging service makes the company (QQ) the largest blog site provider in the survey, surpassing the likes of Windows Live Spaces, Blogger and MySpace.

Following is a report on number of web sites found for each server product found on this web page.

Netcraft Web Server Survey - http://survey.netcraft.com/Reports/200902/

Just Start Pitching: The infallible Sales Pitch

2009-02-25T11:12:00.002+05:30

I recently made a trip to two electronics stores that also sold Laptops. The first is called X-cite the second ‘House of Laptops’. I was looking for a ThinkPad, but both these places did not have the Lenovo ThinkPad with them. X-cite and ‘House of Laptops’ have very marked difference in which they approach customers.

At X-cite, when I didn’t find the ThinkPad the sales guy approached me with the question – “maybe I can help you find some other laptop; if you can only tell me your specifications, features and cost preferences”. I thought about it for a few seconds, and then I left the store.

At ‘House of Laptops’, when I didn’t find the ThinkPad the sales guy immediately started pitching me a new model from Toshiba. I have never bought a Toshiba laptop and there was very little likelihood that I would buy a Toshiba. However, the sales guy totally overwhelmed me with the features of the slick Toshiba model. It had the latest Core 2 Duo, and all the gizmos that you can ask for. It was 1.8kg and priced at Rs. 49,000, with some freebies. And it had spill proof keyboard.

The first guy’s approach was probably more scientific, in terms of collecting customer requirements and then providing the customer with a solution. However, the customer walked off. The second guy just started pitching, without understanding the customer's requirements. Obviously the second approach is much better, there are much more chances of selling a laptop with the second approach. I think there is learning in this.

What do you think?

Contextually Relevant Ads from Google AdSense that made me click

2008-12-23T18:40:00.004+05:30

Today I found an ad that I readily clicked. This is an ad about weekend trip to Lonavala. Such a relevant ad and such great timing. I clicked on the ad, and it actually kept it's word of the pricing at Rs. 3568. Good job Google AdSense!

Following is the contextual ad. I did a little a little more digging, and found 2 more interesting things:

Google AdSense click URL shows up in Chrome (on mouseover), that doesn't show up in any other browser - kind of cool
Google AdSense URL is now pointing to http://googleads.g.doubleclick.net/pagead/ , while earlier it used to point to http://pagead2.googlesyndication.com/pagead/ . Interesting. Obvious integration with DoubleClick; but is that some kind of 'lead generation' ?

Following is the contextual ad.

Some fun with Google Hindi translation!

2008-12-12T10:49:00.003+05:30

Google translation for Hindi is pretty cool. I test drove it, then I got some ideas, here they are, have fun:

English: Google translation rocks!
Hindi: Google अनुवाद चट्टानों!

English: Google translation is cool!
Hindi: Google अनुवाद ठंडा है!

English: Google translation is a kick-ass product!
Hindi: Google अनुवाद एक लात-गधा उत्पाद है!

English: Google translation is a super awesome product!
Hindi: Google अनुवाद एक सुपर भययोग्य उत्पाद है!

Click here to try more stuff.

Yahoo! Email Ads still can’t get geo-targeting right

2008-12-09T10:04:00.001+05:30

Yahoo! Email continues to serve me US ads, wasting money on those high eCPM expandable rich media ads.

What a pity. Yahoo! knows that I live in Pune, India; it can tell the temperature of the city, but cannot geo-target the right ads for me.

There are so many other ad-servers that get it right, why can’t Yahoo! get it right?

Thoughts?

Komli Launches ViziSense: India’s First Free and Open Online Audience Measurement Platform

2008-12-02T14:00:00.002+05:30

I’m very excited today to announce the launch of Komli ViziSense, the first free and open audience measurement platform for India that accurately reports details of site visitor demographics and other audience characteristics. ViziSense helps publishers with an independent measurement system, enables advertisers to access and understand their online audience with precision, and allows ViziSense agencies to plan better media buys.

Read more details here and here.

Some screen shots below:

5 tips for startups for handling the power situation

2008-11-22T15:39:00.002+05:30

Recently we have had a few power outages in our office at Pune. After about 3 days of facing the crisis, I had to find a solution for this problem. Following are 5 tips for startups for handling the power situation. These tips are specifically for early stage startups.

USE LAPTOPS INSTEAD OF DESKTOPS – this is pretty much a no brainer. Laptops consume about 100 watts of power, while desktops consume 500 watts. Apart from the other obvious benefits I recently found out that – you can pretty much lie down on a bean-bag and still work on a laptop. This was after we bought a few bean bags in the office.
USE VIRTUAL MACHINES – this is one of the coolest ideas. Instead of buying 4 low powered machines buy one powerful machine and make 4 virtual machines out of it. So, an example configuration would be 2 x Quad-Core Intel® Xeon® processor E5410 + 16GB RAM + 1TB HDD – split it 4-ways to make into 4 powerful 2-core servers. Pretty awesome, 4 physical machines would consume 2000W power, while the 2xXeon would consume about 500-600W.
INSTALL A “TRUE ONLINE SINEUPS” – these are much more powerful than regular inverter/UPS. They are costly, yet worth every penny. This has much more battery power than a regular inverter/UPS, and can last much longer. They come in 6KVA, 8KVA and 10KVA, on the smaller side.
SEPARATE OUT THE BACKUP FOR SERVERS, laptops and fans/lights - The problem with that is – it adds another point of failure because of more failing parts. You don’t want to hear UPS’ beeping all day in your office.
MOVE TO A FULLY FURNISHED OFFICE that has generator backup – These cost upwards of Rs. 70 per sqft. Pretty high cost for a startup.

To summarize, tips 1, 2 and 3 are the really feasible options.

Got more ideas? Have questions? Send me a message on facebook or Twitter or send me an email at mukul dot kumar at pubmatic dot com .

The Real Cost of Amazon CloudFront

2008-11-19T16:22:00.002+05:30

Amazon introduced web service for content delivery, called Amazon CloudFront yesterday. CloudFront is thought to be a bring a pricing war between the current CDN providers.

If you do a little bit calculations for the real cost of the CDN, it turns out that it is much higher than the advertised pricing, for smaller files.

Following is the effective cost per GB for USA locations for CloudBurst. For example - if your files are 5KB in size, you will actually pay $0.3797 per GB not $0.17 . If your file size is 10MB, then you will pay the advertised price of $0.17 per GB. So, essentially if you are distributing images or movies, CloudFront will be cost effective, however if you are distributing JavaScripts of small size, you may be paying a lot more.

File Size (KB)	Effective Cost per GB
5KB	$0.3797 (123% more)
10KB	$0.2749
20KB	$0.2224
50KB	$0.1910
100KB	$0.1805
500KB	$0.1721
1MB	$0.1710
5MB	$0.1702
10MB	$0.1701
100MB	$0.1700
1GB	$0.1700

Here is how I calculated this - for US locations, data transfer rate (for first 10 TB / month) is $0.17 and request rate is $0.01 per 10,000 requests.
Effective cost per GB = $0.17+(1024*1024/file_size * 0.01/10000);

The world’s most super-designed data center

2008-11-15T09:43:00.002+05:30

Royal Pingdom reported: The newly opened high-security data center run by one of Sweden’s largest ISPs, located in an old nuclear bunker deep below the bedrock of Stockholm city, sealed off from the world by entrance doors 40 cm thick (almost 16 inches).

Facts about the data center:

Originally a nuclear bunker
Located in central Stockholm below 30 meters (almost 100 ft) of bedrock
Can withstand a hydrogen bomb
Houses the Network Operations Center for one of Sweden’s largest ISPs
German submarine engines for backup power
1.5 megawatt of cooling for the servers
Triple redundancy Internet backbone access
Work environment with simulated daylight and greenhouses

Following is one of the photographs, courtsey Royal Pingdom:

Read more here.

VMware Wants to Bring Virtualization to Smart Phones - BUT where are the use-cases

2008-11-10T23:21:00.002+05:30

eWeek.com reports that VMware is going to release a virtualization platform for mobile phones, but gives no real use-case for such a platform. Later in this article they talk about the use-case of creating different profiles in different VMs; isn't that possible today using different profiles on a Windows based operating system, why would anybody create a separate virtual machine to create a different profile? I could imagine (hypothetically speaking) running an iPhone inside a RIM phone, or running an S60 app inside an iPhone - but is that a real use case.

Here are parts of the article:

"VMware is looking to bring its virtualization technology to smart phones and cell phones in 2009 through a new virtualization platform called the VMware Mobile Virtualization Platform, or MVP. The platform will use a small-footprint hypervisor that will allow users to have multiple virtual machines on their smart phones in the same way a desktop or notebook can host different virtual environments. Mobile Virtualization Platform, or MVP, which consists of a small, bare-metal hypervisor— 20KB to 30KB —that will work with a number of mobile devices based on an ARM processor.

“This virtualization layer that we have is just like the one on the server and desktops, and it will allow customers to run multiple virtual environments on the phone,” said Krishnamurti. “We think there are some interesting use cases. One is that many people have one phone for work and another is a personal phone. With virtualization, you can have one device that runs both environments in two isolated virtual machines. The work profile and the personal profile are completely separated.”

Right now, the VMware MVP platform will support a number of mobile devices based on Linux, Windows CE and Symbian, which is now owned by Nokia. Later, Krishnamurti said, VMware will add support for Google’s Android operating system."

Cloud Computing: How important is "data locality" from a costing perspective?

2008-11-07T10:59:00.004+05:30

Nicholas Carr wrote an excellent article about cloud computing "The new economics of computing".

"In late 2007, the New York Times faced a challenge. It wanted to make available over the web its entire archive of articles, 11 million in all, dating back to 1851. It had already scanned all the articles, producing a huge, four-terabyte pile of images in TIFF format. But because TIFFs are poorly suited to online distribution, and because a single article often comprised many TIFFs, the Times needed to translate that four-terabyte pile of TIFFs into more web-friendly PDF files.

Working alone, he uploaded the four terabytes of TIFF data into Amazon's Simple Storage Service (S3) utility, and he hacked together some code for EC2 that would, as he later described in a blog post, "pull all the parts that make up an article out of S3, generate a PDF from them and store the PDF back in S3." He then rented 100 virtual computers through EC2 and ran the data through them. In less than 24 hours, he had his 11 million PDFs, all stored neatly in S3 and ready to be served up to visitors to the Times site.

The total cost for the computing job? Gottfrid told me that the entire EC2 bill came to $240. (That's 10 cents per computer-hour times 100 computers times 24 hours; there were no bandwidth charges since all the data transfers took place within Amazon's system - from S3 to EC2 and back.)"

One thing missed in the "NYT TIFF to PDF conversion computational task" is the mention of data transfer cost of uploading 4TB TIFF images into S3.

Doing some simple computations – Amazon would charge about $409.60 for uploading 4TB data into S3, and would charge an additional $261.12 for downloading the processed PDF files, which were 1.5TB in size. That is about $670.72. In addition there will be bandwidth charges of this 5.5TB data transfer from the NYT datacenter, 4TB out and 1.5TB in, I am sure that will be of the order of $400-$600. That could take the data transfer costs to $1000-$1200 range.

In addition to that – consider the amount of time it would take to transfer such a data. At 10Mbps, it would take 53.4 days to transfer this data.

Using Hadoop on EC2 is definitely a great idea, and is very helpful, however the locality of data also matters a lot. Moving data, in my opinion costs a lot, and sometimes undermines the computational costs ($240 here).

Let me know your thoughts.

Finding and fixing memory issues using Valgrind with example of Apache+FastCGI web application

2008-11-02T01:08:00.001+05:30

My colleague, Anand Das, has written an excellent article on using Valgrind to debug live web applications. Read the full article here.

Is ‘Divide and Conquer Algorithm’ a human instinct?

2008-10-11T19:36:00.002+05:30

Last Friday I was with my son at a restaurant, I was browsing the menu, when I thought I will give my son a quick problem, and see how far he reaches in solving it. The menu had 5 pages, each page had about 20 items.

I asked my son if he could find the ‘cheapest’ and the ‘costliest’ items in the entire menu.

My son is 6 years old, and does some very basic mathematics at the school, such as arranging numbers in ascending and descending orders, simple single-digit additions and subtractions, etc.

I was amazed to find that, my son quickly found the cheapest item on page #1, then the cheapest item on page #2, chose the lesser of two ( MIN(page#1,page#2) ), and then recursively did that for pages 3, 4 and 5. And gave me the cheapest item on the menu (Rs. 6). He then did the same for costliest item, he found the costliest item on page #1, then the costliest item on page #2, chose the higher of two ( MAX(page#1,page#2) ), and gave the answer (Rs. 95). [We were at a Udipi restaurant, so things are not that costly here.]

I never taught him the ‘divide and conquer algorithm’, in fact I have never taught him any algorithm.

Is ‘Divide and Conquer Algorithm’ a human instinct? Does it come without learning?

Thoughts?

5 Hacks for Startup Hiring in India

2008-10-06T16:18:00.002+05:30

Here are some thoughts on hiring for a Startup in India. My experience with hiring in India for the last fifteen years, in one word, has been – “Awesome” ! In Pune I have met some of the best Programmers and Designers in the world and work with many of them. There are some of most hardworking, smart and knowledgeable individuals, who love to crank code (read an interesting post here, under the “People” section). I love working with the great guys at Komli!

Hiring in India is different than hiring in other parts of the world. The following thoughts are written for an employer in mind, especially a startup employer. These thoughts are in a random order, and based on personal experiences. Please don’t equate my consistent use of ‘he/him’ with a gender bias.

1. “Offer acceptance is not equal to JOINING”

This is something you learn the hard way. It is very difficult to believe that a candidate talks so nicely, and accepts your job offer, only to NOT show up on the joining date. This is a shocker, which takes several days to recover from. If the candidate is good he calls up/e-mails you a few days in advance, telling you that he cannot join. Many will not inform you, and simply won’t show up on the joining date.

My recent experience – “a senior managerial candidate, who was relocating from the USA to India, accepted the offer after several negotiations that went on for weeks. He was very happy and I was very happy, that we have a deal. The day he was supposed to land in Pune, and join after a few hours of landing – he didn’t show up. I patiently waited until the evening, and next morning. Emailed him, and found out that instead of flying into Pune, he landed into Bangalore, and joined another large company yesterday. How nice.”

Accept the fact – a hire is only a “probability” until the day he shows up. This probability increases as the date of joining comes near. An offer acceptance on e-mail, or in hardcopy are still probabilities of joining.

The way I would handle this is – a) don’t count on a hire until he joins, b) always plan for backups – no hiring is complete until the last guy joins, c) keep calling the candidate every few days, to find out if he is going to join – if he tells that he is not joining, it’s better to know that early on, rather than on the last date, and d) if the numbers are large – over-hire, to compensate for the probability.

2. “The Resume”

I have found that many resumes have inaccurate information in them. You can actually build a “probability-framework” on what percentage of a resume is true – based on some of the key parameters of the resume – such as skill-set (Java, PHP, RoR, ASP.NET, C/C++, C#, etc). Try that, it works.

The way I handle this is – talk to the candidate, find out what he has done. Correlate that with his resume. Most of the times they match.

Another interesting parameter is the – “keyword density”. My personal experience has been that the higher the keyword density, the more likely is that the candidate is bogus. You cannot learn – all of “C, C++, Java, PHP, MySQL, Oracle, OLTP, Apache, Tomcat, RoR” in 2 years :-)

3. “The Sourcing”

Sourcing of resumes has a major impact on the success rate. I think it is very important to access the success rate of each of the source of resumes – direct, referral, recruiters, newspapers, online portals, etc. You may be surprised to know that there may be a difference of 10x in the conversion rates of each of these sources – so you should focus on the source that has the highest conversion rate.

For startups referrals work the best. Keep your employees happy, so that they find more friends who want to become happy!

4. “The Interview”

A few things at the top of my head are following:
Do initial screenings before you go too deep into technical discussion. If the candidate is not good, let’s find out in the first ten minutes of discussion, so that you save time on both sides. One important thing in my mind is – ask questions about your most recent problem that your company is facing, find out if he can solve that problem or not. Even if a guy can solve the most complex algorithm problems, or he can do the most optimal data structure design – can he solve your current (or past two-three) problems? Make sure you factor that into the overall decision. Don’t compare the candidate to yourself – “he is not like me; I can do it better than him”. It is very difficult to find a guy better that yourself, don’t try that :-)

5. “The Timing”

Try to keep “good” interviews at the top of the day, during mornings. You are in the office at 9AM, if the candidate doesn’t show up, or doesn’t pickup the phone – that does very bad things to your day. It’s a difficult thing to do, but I try and keep most interviews at the later part of the day.

Well, that’s all I have for now. There are many more things, but I wanted to keep it simple.

Got any more ideas, send me a message on facebook or Twitter?

Why doesn't Google Chrome render my bullets right?

2008-09-17T11:38:00.001+05:30

Why doesn't Google Chrome render my bullets right?

See the rendering of bullets using FF, IE7 and Chrome below:

EOS 5D Mark II Digital SLR camera - Is this the new future proof SLR Camera?

2008-09-17T11:20:00.003+05:30

Is this the new future proof SLR Camera?

EOS 5D Mark II Digital SLR camera, the long-awaited successor to Canon's highly popular EOS 5D, introduced in 2005 [original post here].

21.1-megapixel full frame, 24 x 36mm CMOS sensor, DIGIC 4 imaging processor
Expanded sensitivity range from ISO 50 to ISO 25,600
16:9 Full HD video capture at 1920 x 1080 pixels and 30 fps as well as 4:3 standard TV quality (SD) video capture at 640 x 480 pixels and 30 fps

Read more here.

PubMatic interview with Amar Goel

2008-09-17T11:03:00.004+05:30

PubMatic interview with Amar Goel, CEO PubMatic/Komli:
Amar talks about ad optimization, how PubMatic benefits publishers, and how ad-price index shows how adprices are trending:

Chrome shine for Google

2008-09-11T17:25:00.002+05:30

I wrote a review of Google Chrome Browser for Financial Express. It appeared on page-8 of the newspaper, and is also available online at - http://bit.ly/chromefe .

Check it out.

Google Chrome Tested

2008-09-03T18:53:00.001+05:30

I have been testing Google Chrome for last 6 hours now. Here is a short review. Overall I don’t find anything overwhelmingly amazing that would make me WOW! The Omnibox is cool, but I feel that it makes a marginal enhancement in my browsing experience. The most important thing I noticed is that – Chrome is slow; it’s slower than FireFox. I have to admit that I am so used to Firebug, that I almost view the Net-element of Firebug every few minutes, I find such a functionality missing in Chrome. I like the Chrome->Developer->Debug Javascript, but it has marginal stuff that I need. Also, check this – in Chrome, click on Developer menu, and try to shift-Windows-tabs (by alt-tab), that doesn’t work. Why has Chrome disabled my Windows-alt-tab switch? Web site seem to work normally, rendering and JavaScript seems to be working fine.

I feel that Javascript execution is very very fast. I tested a page where FF normally give a Javascript-timeout (stop, continue …), Google Chrome ran just fine and delivered me the Javascript rendering in less time than expected. I will therefore need to do more Javascript testing on Chrome.

Though I am a little concerned about the vulnerability discussed here.

More later.

Social network popularity around the world

2008-08-12T20:25:00.002+05:30

Very interesting data on popularity of social networks posted on this web site. The social networks we included in this survey were MySpace, Facebook, Hi5, Friendster, LinkedIn, Orkut, Last.fm, LiveJournal, Xanga, Bebo, Imeem and Twitter.

Facebook is most popular in Turkey and Canada.
Friendster and Imeem are most popular in the Philippines.
LinkedIn is most popular in India.
Twitter is most popular in Japan.
LiveJournal is more popular in Russia than it is in the United States.
Orkut is more popular in Iran (10th country popularity-wise) than it is in the United States.
MySpace is the only social network which is most popular in the United States.

Read more here.

Servers: What do they really cost?

2008-08-12T01:25:00.001+05:30

Really interesting post on Forbes.com about how much a server really costs, some excerpts are following:

Spending $2,500 on a server really means spending between $8,300 and $15,400 in facility capital to provide the necessary space for housing the server and powering it.

A single rack of blade servers will consume $1 million in facility capital. Pretty staggering! And this is the underlying source of the boom in data-center construction driven by the need to house 5 million additional servers.

Read more here.

Cloud Availability

2008-07-27T12:04:00.006+05:30

Cloud Computing has become very widespread with startups as well as divisions of banks, pharmaceuticals companies and other large corporations using them for computing and storage.
Amazon Web Services has led the pack with it's innovation and execution, with services such S3 storage service, EC2 compute cloud, and SimpleDB online database.

Many options exist today for cloud services, for hosting, storage and application hosting. Some examples are below:

Hosting	Storage	Applications
Amazon EC2	Amazon S3	opSource
MOSSO	Nirvanix	Google Apps
GoGrid	Microsoft Mesh	Salesforce.com
AppNexus	EMC Mozy
Google AppEngine	MOSSO CloudFS
flexiscale

[A good compilation of cloud computing is here, with a nice list of providers here. Also worth checking out is this post.]

The high availability of these cloud services becomes more important with some of these companies relying on these services for their critical infrastructure. Recent outages of Amazon S3 (here and here) have raised some important questions such as this - S3 Outage Highlights Fragility of Web Services and this.

[A simple search on search.twitter.com can tell you things that you won't find on web pages. Check it out with this search, this and this.]

There has been some discussion on the high availability of cloud services and some possible solutions. For example the following posts - "Strategy: Front S3 with a Caching Proxy" and "Responding to Amazon's S3 outage".

Here I am writing of some thoughts on how these cloud services can be made highly available, by following the traditional path of redundancy.

The traditional way of using AWS S3 is to use it with AWS EC2 (config #0).

Configurations such as on the left can be made to make your computing and storage not dependent on the same service provider.
Config #1, config #2 and config #3 mix and match some of the more flexible computing services with storage services.
In theory the compute and the storage can be separately replaced by a colo service.

The configurations on the right are examples of providing high availability by making a "hot-standby".

Config #4 makes the storage service hot-standby and config #5 separates the web-service layer from the application layer,
and makes the whole application+storage layer as hot-standby.

A hot-standby requires three things to be configured - rsync, monitoring and switchover.

rsync needs to be configured between hot-standby servers, to make sure that most of the application and data components
are up to date on the online-server. So for example in config #4 one has to rsync 'Amazon S3' to 'Nirvanix' - that's pretty
easy to setup. In fact, if we add more automation, we can "turn-off" a standby server after making sure that the
data-source is synced up. Though that assumes that the server provisioning time is an acceptable downtime,
i.e. the RTO (Recovery time objective) is within acceptable limits.

This also requires that you are monitoring each of the web services. One might have to do service-heartbeating -
this has to be designed for the application, this has to be designed differently for monitoring Tomcat, MySQL,
Apache or their sub-components. In theory it would be nice if a cloud computing service would export APIs,
for example an API for http://status.aws.amazon.com/ , http://status.mosso.com/ or http://heartbeat.skype.com/.
However, most of the times the status page is updated much later after the service goes down. So, that wouldn't help much.

Switchover from the online-server/service to the hot-standby would probably have to be done by hand.
This requires a handshake with the upper layer so that requests stop and start going to the new service
when you trigger the switchover. This might become interesting with stateful-services and also where
you cannot drop any packets, so quiscing may have to be done for the requests before the switchover takes place.

Above are two configurations of multi-tiered web-services, where each service is built on a different cloud service. This is a theoretical configuration, since I don't know of many good cloud services, there are only a few. But this may represent a possible future, where the space
becomes fragmented, with many service providers.

Config #7 is config #6 with hot-standby for each of the service layers.
Again this is a theoretical configuration.

Cost Impact
Any of the hot-standby configurations would have cost impact - adding any extra layer of high-availability immediately adds to the cost, at least doubling the cost of the infrastructure. This cost increase can be reduced by making only those parts of your infrastructure highly-available that affect your business the most. It depends on how much business impact does a downtime cause, and therefore how much money can be spent on the infrastructure.

One of the ways to make the configurations more cost effective is to make them active-active configuration also called a load balanced configuration - these configurations would make use of all the allocated resources and would send traffic to both the servers. This configuration is much more difficult to design - for example if you put the hot-standby-storage in active-active configuration then every "write" (DB insert) must go to both the storage-servers, writes (DB insert) must not complete on any replicas (also called mirrored write consistency).

Cloud Computing becoming mainstream
As cloud computing becomes more mainstream - larger web companies may start using these services, they may put a part of their infrastructure on a compute cloud. For example, I can imagine a cloud dedicated for "data mining" being used by several companies, these may have servers with large HDDs and memory and may specialize in cluster software such as Hadoop.

Lastly I would like to cover my favorite topic -why would I still use services that cost more for my core services instead of using cloud computing?

The most important reason would be 24x7 support. Hosting providers such as servepath and rackspace provide support. When I give a call to the support at 2PM India time, they have a support guy picking up my calls – that’s a great thing. Believe me 24x7 support is a very difficult thing to do.
These hosting providers give me more configurability for RAM/disk/CPU
I can have more control over the network and storage topology of my infrastructure
Point #2 above can give me consistent throughput and latency for I/O access, and network access
These services give me better SLAs
Security

Comparing Clouds: Amazon EC2, Google, AppNexus, and GoGrid

2008-07-23T11:36:00.002+05:30

InfoWorld has an awesome article published by Peter Wayner, who compares various cloud computing services - Amazon, Google, AppNexus, and GoGrid. Read more here.

Following is an excerpt:

Scientists found a workable way of reducing CO2 levels

2008-07-22T13:43:00.004+05:30

Physorg.com reports - Scientists say they have found a workable way of reducing CO₂ levels in the atmosphere by adding lime to seawater. And they think it has the potential to dramatically reverse CO₂ accumulation in the atmosphere, reports Cath O'Driscoll in SCI's Chemistry & Industry magazine published today.

Adding lime to seawater increases alkalinity, boosting seawater's ability to absorb CO2 from air and reducing the tendency to release it back again.

The important point is that when you put lime into seawater it absorbs almost twice as much carbon dioxide as is produced by the breaking down of the limestone in the first place.

The Cquestrate project has a web page; here's how they describe the idea; and here's a chance of getting involved. I think this is HUGE!

They also have this awesome slideshow presentation, which goes into the gory but interesting details of all the chemical reactions.

Here is their idea of how they would do it (subject to economic feasibility):

The Nullarbor Plain is the world's largest single piece of limestone and occupies an areas of about 200,000 km²
With an average thickness of 50m, there is 10,000 km³ of limestone
To sequester 1 billion tonnes of carbon (GtC) would require the excavation of 1.5 km³ of limestone
Between 1750 and 2004 humankind has emitted 305 GtC. Current emission rates are about 7 GtC per year
Thus employing this process on 500 km³ of limestone (about 5% of the limestone in the Nullarbor Plain) would return carbon dioxide levels to pre-industrial levels.

I hope all of this comes true. I think, economic viability of converting CaCO3 into CaO will be the biggest hurdle.

Early YouTube Engineer talks about outages, system scalability

2008-07-13T12:02:00.007+05:30

Gigaom.com has posted a very interesting presentation by Cuong Do, an early software engineer who’s now manager of the site’s Core Product Engineering group.

Gigaom.com has a video, here is the transcript of that video.

Introduction

Do’s talk was titled “Behind the Scenes: A Look Into YouTube’s Infrastructure,” had harrowing tales of outages; gory details about the specific languages, architectures, and tools YouTube uses. “One of the key phrases we had in the early days was ‘These are good problems to have,’” Do said. “And after a while we’re like, ‘I’m going to kill the next person who says that.’”

Do describes the “Early team”

2 sys admin
2 scalability software architects
2 feature developers
2 network engineers
1 DB admin
Zero chefs

Algorithm for handling rapid growth

while (true) {
identify_and_fix_bottlenecks();
drink();
sleep(TOO_LITTLE);
notice_new_bottleneck();
}

Web Request flow

End user browser
NetScalar load balancer
Web Servers (bank of)
Apache
Local app server
Python app
Memcached
DB

Video served through

CDN - internal project from Google (for non-US content)
Video – colo servers – from various US locations

Key technologies

Linux (SuSE 10.x)
Apache (2.2.x) / lighthttpd 1.4.x
lighthttpd - very fast in handling large files
MySQL 5.0.x - metadata storage
Python
[Very difficult to recruit people for a small company, Faster to get 10 machines, than 1 dev, so got more machines that run slow code.]
Google technologies - search, Bigtable (video thumbnails), GFS (internal storage, transfer buffers)

Started with 1 DB server

Replica for backup, replica for reporting (later)
Vertical partitions for DB, takes part of the web site (more later)
Multiple users for DB, for scalability
Associate one user with one partition

Scalability challenges

Rapid unpredictable growth (user growth always exceeds any amount of hardware scale)
Passionate users (users who video-blog their life)
New features (recommendation algorithms, compatibility, social graphs – always blew off scalability predictions)
Pushing hardware and software boundaries (if you are running a hardware or software too close to its limits, it's more likely to fail)
Unknown problems (issues that you don't find on Google search, or issues that even the vendor of the 3rd-party-software doesn’t know)

One example of an issue we had

Subject: oh @#!%
Date: October 22, 2005 2:24:33 AM PDT
We can't accept any more videos, too many videos.

All thumbnails were stored in separate sub-directories, and all sub-directories were in one flat directory.
More than 10K files in a directory, out of Inodes problem.
Wasn't too difficult to solve, went to a tree structure.
Wasn't the funniest thing to do, with all videos being uploaded

YouTube: 5 hours Outage

MySQL gave error, checksum has failed
Checksum stored for every page (15K of data)
MySQL checksum failed, it puked, lost 4-5 hours of data
Took 4-5 hours to recover
We found lots of questions, but no answers for this problem
"Maybe this is a hardware issue, not everybody is having it"
Found exactly one-blog-post, turns out the combination of the RAID card and the other I/O card, can sometimes cause interesting fluctuations in voltage
So the data was fine in the disk, the voltage fluctuation caused the CPU to read different data
It took weeks to figure this out

One of most favorite ones: Again ran out of disk space!!

Awesome.

Best Web 2.0 Sites in India

2008-07-10T17:58:00.005+05:30

I am compiling a list of best Web 2.0 sites in India. I am most likely going to do a writeup on each of the sites, and then suggest making a group of like minded people, who can exchange great ideas.The initial list is following in no particular order. Feel free to add more in this list, by commenting this post:

http://www.ApnaBill.com
http://bumsonthesaddle.com/
http://quillpad.com/
http://videoontherocks.indya.com/ (gave me a bunch of "[MySQL][ODBC 3.51 Driver]Can't connect to MySQL server on '192.168.1.5' (10061)" errors)
http://www.cleartrip.com/
http://bigflix.com/
http://www.cleartrip.com/
http://www.magicbricks.com/
http://www.yulop.com/
http://bookeazy.com/
http://iBibo.com/
http://Burrp.com/

Google Advanced Search - now with dates!

2008-07-07T11:38:00.002+05:30

Nice:

Designing Caches for Highly Scalable Web 2.0 Applications

2008-06-28T21:41:00.000+05:30

Unix/Linux file systems have been designed in a way that reads are heavily cached and sometimes pre-fetched. There are various techniques, algorithms and methods for read caching, and each file system has its somewhat unique method and therefore performance. Most file systems would use page-cache for caching read-I/O and buffer cache for caching the metadata.

There has been immense amount of research in this area – of how to improve the read performance using caching (see here and here).

Enter highly scalable Web 2.0 era, enter Facebook - if you look at the Facebook IO-profile in my previous post – 92% of the read (for photos) are served by the CDN. What that means is reads will only happen once, after that the file will be cached in the CDN and the read will never go to the backend storage (NetApp filer in this case). So all the file system caching is probably going waste, since we are never going to read from the file-system-cache ever. Facebook photos are cached in CDN for 4.24 years (their http cache-control max-age is 133,721,540), which means the CDN will not go back to the origin server for that period.

This raises interesting questions – do file systems really need to do any caching, what is the read-write ratio for such an application, how can this file system be better tuned for such an application?
Can file system cache be better used for pre-fetching the entire metadata in the cache, so that Facebook NetApp filer has to do fewer than 3 reads for reading a photo?

Thoughts?

Facebook Photo Storage Architecture

2008-06-27T11:39:00.003+05:30

Awesome presentation called "Facebook - Needle in a Haystack: Efficient Storage of Billions of Photos", here is an excerpt. You should see the full presentation (if you can get flowgram to work).

Facebook uses MySQL, Memcache, Apache, PHP and Extensions in their application stack.

Facebook uses NetAPP filers for storing files.

Facebook scale of photos

~6.5 billion total images, 4-5 sizes stores for each image => ~30 billion files => 540TB total storage capacity.
~475,000 images server per second at peak – most through CDNs.
~100 million uploaded per week.

Facebook uses a 4-tier architecture for serving profiles and photos.

The first tier is CDN, then their proprietary “Cachr”, then their photo servers then the NetApp filers.

Cachr

“Protects the origin for profile pictures
Based on modified evhttp
Uses memcache as backing store

Microsecond response time on cache hit
Server can die or restart without losing cache

File handle cache

Based on lighthttpd
Uses memcache as backing store
Reduces metadata workload on NTAP

NetApp storage architectural issues

NetApp Storage is overwhelmed with metadata
~3 disk reads to to read one photo
Totally bottlenecked on disk bandwidth

Thus, heavy reliance on expensive CDNs to serve reads:

99.8% hit rate in CDN for profile images
~92% hit rate for photos
Drastically reduces load on the storage

All models are wrong, and increasingly you can succeed without them.

2008-06-26T14:56:00.004+05:30

Says Peter Norvig, Google's research director, as an update to George Box's maxim.

This is an awesome article about Petabyte size datasets and why correlation of data is enough, instead of finding a reason why datasets are related and building a model around it. Read some excerpts here, and full article via this link.

Excerpts from the original article:
Sixty years ago, digital computers made information readable. Twenty years ago, the Internet made it reachable. Ten years ago, the first search engine crawlers made it a single database. Now Google and like-minded companies are sifting through the most measured age in history, treating this massive corpus as a laboratory of the human condition. They are the children of the Petabyte Age.

Google's founding philosophy is that we don't know why this page is better than that one: If the statistics of incoming links say it is, that's good enough. No semantic or causal analysis is required. That's why Google can translate languages without actually "knowing" them (given equal corpus data, Google can translate Klingon into Farsi as easily as it can translate French into German). And why it can match ads to content without any knowledge or assumptions about the ads or the content.

This is a world where massive amounts of data and applied mathematics replace every other tool that might be brought to bear. Out with every theory of human behavior, from linguistics to sociology. Forget taxonomy, ontology, and psychology. Who knows why people do what they do? The point is they do it, and we can track and measure it with unprecedented fidelity. With enough data, the numbers speak for themselves.

If the words "discover a new species" call to mind Darwin and drawings of finches, you may be stuck in the old way of doing science. Venter can tell you almost nothing about the species he found. He doesn't know what they look like, how they live, or much of anything else about their morphology. He doesn't even have their entire genome. All he has is a statistical blip — a unique sequence that, being unlike any other sequence in the database, must represent a new species.

Petabytes allow us to say: "Correlation is enough."

3 awesome resources for Cross Browser Coding

2008-06-20T17:24:00.005+05:30

I have been doing a bunch of reading, here are three of best articles I have read on cross browser coding. They cover – a) rendering, b) events compatibility and c) cookie restrictions.

Rendering - How to get Cross Browser Compatibility Every Time.

Here is a quick summary for those of you who don't want to read the whole article:

Always use strict doctype and standards-compliant HTML/CSS
Always use a reset at the start of your css
Use opacity:0.99 on text elements to clean up rendering in Safari
Never resize images in the CSS or HTML
Check font rendering in every browser. Don't use Lucida
Size text as a % in the body, and as em's throughout
All layout divs that are floated should include display:inline and overflow:hidden
Containers should have overflow:auto and trigger hasLayout via a width or height
Don't use any fancy CSS3 selectors
Don't use transparent PNG's unless you have loaded the alpha

Events compatibility table. This article compares the event handling for IE 5.5, IE 6, IE 7, IE8b1, FF 2, FF 3b5, Saf 3.0 Win, Saf 3.1 Win, Opera 9.26, Opera 9.5b and Konqueror 3.5.7. Very detailed article. Must read for all Javascript programmers.

Cookie Restrictions - Browser cookie restrictions very nicely documented here. An excerpt follows:

Microsoft indicated that Internet Explorer 8 increased the cookie limit per domain to 50 cookies but I’ve found that IE7 also allows 50 cookies per domain. Granted, this may have been increased with a system patch rather than having the browser’s first version ship like this, but it’s still more than the 20 that was commonly understood to be the limit.
Firefox has a per-domain cookie limit of 50 cookies.
Opera has a per-domain cookie limit of 30 cookies.
Safari/WebKit is the most interesting of all as it appears to have no perceivable limitCookie header. The problem is that the header size exceeded the limit that the server could process, so an error occurred. through Safari 3.1. I tested setting up to 10,000 cookies and all of them were set and sent along in the

Selling a Used Car in India - eBay.in, CarWale.com, Kijiji.in compared

2008-06-13T12:42:00.003+05:30

I put my car-ad on eBay.in (http://motors.ebay.in/), CarWale.com and Kijiji (http://www.kijiji.in/).

The good news is that I could sell my car in less than 2 weeks after putting the ad, so I am happy.

Here is a short review of the sites:

eBay.in has a very cool interface, trusted site, many cars are sold every hour on eBay.in .I had a small issue with updating my classified-ad, and their support emailed me within a few hours. That's great support. They charge only Rs. 30 for placing a classified ad.

CarWale.com also has a great interface. The best thing I liked about CarWale.com is that whenever there is a new lead, CarWale.com would send me an SMS on my mobile telling me the name and the phone number of the person interested – that comes in extremely handy, since I can call the person immediately, and don’t have to get online to see the person’s mobile number. The second best thing about CarWale.com is that when I started to fill in the expected price of the car, they showed a nice popup suggesting me the resale-value of this car for that year, etc. That was really cool. Though the prices were approximately 20% inflated, so it initially raised my expectations a lot about the resale value. They charged me Rs. 500 for putting the ad.

Kijiji.in is cool too. Interface is cool; they sent me 2 emails on new leads – the prices were pathetically low. I never visited their web-page to find out. Kijiji hasn’t charged me any money so far for putting the ad.

CarWale.com generated the most number of leads (6), followed by eBay.in (4), followed by Kijiji.in (1).

I sold my car last week. Overall, I had a great experience.

Web Service for Content Distribution Network

2008-06-12T22:55:00.005+05:30

Having used CDN for a year now, I can say that the complexities of CDN deployment, origin server configuration, Apache configuration, pricing models, file placement, cache flush makes it pretty non-trivial deployment.

Just wondering why not have a "web service model for CDN".

Google recently introduced Google AJAX Libraries API.Google would place some of the popular AJAX libraries on their CDN servers thereby allowing caching, gzip etc. - thereby making the web pages load faster. Great idea! That could make some web pages really fast, quick to load.

Let's make CDN simple. Let's make hosting a file on CDN a 2 step process. See the following mockup:

The entire process of - origin server configuration, Apache configuration, pricing, file placement, cache flush etc. can be made a 2 step process. A wrapper can be written on top of this entire process.

The pricing model can easily cover the cost of the server and the cost of developing such a web service. This could be a service on top of some of the popular CDN providers such as Panther Express, Level3, Limelight Networks, Akamai, etc.

I am surprised Amazon doesn't provide this service in conjunction with EC2 and S3.

Thoughts?

Update: Just read an interesting article - "10 Easy Steps to use Google App Engine as your own CDN". I am going to try it out, and see what the latencies look like, from different cities in the world.

Gartner states the obvious, yet again

2008-06-01T00:35:00.002+05:30

Gartner states the obvious, yet again:

Through TechMeme and 'Business of IT' ...

The good folks over at the Gartner Group have revealed the top 10 technologies that they believe will change the world over the next four years:

Multicore and hybrid processors
Virtualization and fabric computing
Social networks and social software
Cloud computing and cloud/Web platforms
Web mashups
User Interface
Ubiquitous computing
Contextual computing
Augmented reality
Semantics

Give me a break guys. Isn't this obvious.

World's cheapest airbags at $10 a piece

2008-05-31T23:35:00.001+05:30

Autoliv IFB India will develop the world's cheapest airbags for Tata Motors's Nano at $10 a piece. The Nano is likely to have two frontal airbags for the high-end variants. The cheapest airbag which will be made for the Nano will cost about 10 USD, which is around Rs. 400. Autoliv is a global leader in automotive safety components, being the largest airbag manufacturer.

Compare this typical airbag that would cost anywhere between $400 to $1000 (see wikianswers and page 6 of this document).

Way to go Nano and Autoliv!

More news here and here.

Twitter and TinyURL

2008-05-24T18:46:00.008+05:30

Twitter is now probably as ubiquitous as blogging. It’s been used almost as a real-time-blogging-platform. Which is cool – I love Twitter.

If Twitter is ubiquitous – so is TinyURL. TinyURL is a service that will shorten long URLs and make them look good. If you use Twitter, you will know how common it is to find http://tinyurl.com/64f3j9 kind of URLs. TinyURL is using a 6 digit base-36 shortened representation of a URL, which means they can shorten as many as 2,176,782,336 URLs – they claim to shorten 74 million URLs at the moment.

I did a quick research on how many times TinyURL is used on Twitter, and compared it with 12 of it's competitors. Following is a chart compares the number of times a 'shorted-URL-service ' is used on Twitter (it's a logarithmic chart):

I was wondering – How is TinyURL making money?

I see that there are ads on the page, but the ads don’t seem very relevant. In fact on the right side (below the fold) there is a Google-link-unit that shows competitors’ ads, such as "Cloak Urls", "URL Redirecting" and "Smaller URL".

1. It will be cool if TinyURL.com puts ads that are relevant to the previous-page where the user is coming from. Typically, I would browse a page, find the content useful, and then go to tinyurl.com with the intention of making the URL shorter. So, if TinyURL were to show ads relevant to the previous-page where I come from, that’ll be a lot more relevant, and I will be more likely to click on those ads. From a technical standpoint – this is not going to be very easy, but it should be possible.

2. Also, it might be better if TinyURL puts the “donate” links at more relevant places, and maybe put some logic such as – if this is a repeat user, then be more proactive in asking for a donation (modified screenshot attached). If, I was using TinyURL for the 4th time in the day, I would definitely like to donate some money to TinyURL. In the screenshot below, when somebody clicks on "Donate money & Make TinyURL", TinyURL give the options of PayPal and Amazon payments:

Great service – TinyURL + Twitter.

Storage shipments - 1 Billion Gigabytes a year!

2008-05-08T10:18:00.002+05:30

Stephen Lawson reports on Computerworld - Demand for storage is doubling every 18 to 24 months, and within five years, Roberson expects to see a "yottabyte year" when the industry as a whole ships 1 yottabyte (a billion gigabytes), or 1,000 zettabytes, of storage capacity.

HP’s new HP StorageWorks 9100 Extreme Data Storage System (ExDS9100) has a base configuration will consist of four blade servers and three storage blocks, with 246TB of storage. Customers will be able to add either type of capacity independently of the other. With two racks, a system can have as much as 820TB of storage capacity.

The ExDS9100 is scheduled to ship in the fourth quarter. HP predicted that it will cost less than $2 per gigabyte in a typical configuration.

How to - Convert Word Docs to Web Pages

2008-05-07T23:38:00.001+05:30

Useful ...

Convert Word Docs to Web Pages

http://howto.wired.com/wiki/Convert_Word_Docs_to_Web_Pages

Microsoft Word has its place, but that place isn't the web. If you've ever tried to convert a Word document to an HTML document, you know that Word's built-in tools can have disastrous results -- bloated files, proprietary markup and exposed personal information are among the gems you'll get with Word's "Convert to HTML" function.

To get to a semi-sane starting point, try using Word's "Save As: Web Page, Filtered" rather than the regular web page option. This will strip out many of the proprietary tags and won't include potentially personal and revealing info contained in the File Properties dialog.

TinyMCE

Another viable option is TinyMCE, a JavaScript Rich Text Editor that offers a "Paste from Word" option. Paste From Word is intended to be used by those who would like to just "Select All" in Word and paste the content into TinyMCE. Depending on the complexity of your document, TinyMCE may be able to fix some of Word's styling quirks and output usable HTML.

Textism

The good folks over at Textism have a tool that will, to quote the Textism website, "strip Microsoft's proprietary tags and other superfluous noise from Word-generated HTML documents." The results are not only much closer to standards compliant web markup, they also create much much smaller, quickly loading pages.

I used Textism to convert this document from Word to clean HTML. I think, it did a pretty good job.

Performance Showdown - SSDs vs. HDDs

2008-04-29T23:39:00.003+05:30

Very interesting data posted on slashdot:

"Computerworld compared four disks, two popular solid state drives and two Seagate mechanical drives, for read/write performance, bootup speed, CPU utilization and other metrics. The question asked by the reviewer is whether it's worth spending an additional $550 for a SSD in your PC/laptop or to plunk down the extra $1,300 for an SSD-equipped MacBook Air? The answer is a resounding No.

Surprising performance results

“I used HD Tach to test the drives' performance -- and got some interesting results. It was the mechanical Momentus drive (non-SSD) that scored the highest burst speed at 214.3MB/sec. The Crucial SSD came in second at 137.3MB/sec., but the desktop Barracuda (non-SSD) and its 135MB/sec. clung to its heels. Advanced Media's Ridata drive trailed the pack at a leisurely 71.2MB/sec. While the two mechanical drives and the Ridata SSD posted average reads in the 54MB-to-55MB/sec. range, Crucial forged ahead at 120.7MB/sec.

SSDs are highly praised for their boot speed, so I would have been remiss had I relied solely on a standardized test. The results were a bit surprising. Crucial's SSD and the two Seagate devices all required 39 to 40 seconds to cold boot to the desktop. (There are a few minutes of behind-the-scenes activity during a Vista boot, but I determined that the boot was complete once the Windows sidebar appeared.) Ridata did best of them all, with a boot time of 32.1 seconds, although that's hardly the blazing speed you might expect from an electronic versus a mechanical device.”

Moving data

“Finally, because these SSDs have a comparatively small capacity, it's most likely that you will be transferring data from your laptop after a day's work. So I took 4,666 files and folders (a total of 8.05GB) and copied them to the drives and then copied them from those drives. I used the same secondary drive as source and destination in all cases.

Neither of the SSDs fared very well when having data copied to them. Crucial needed 243 seconds and Ridata took 264.5 seconds. That's over four minutes. The Momentus and Barracuda hard drives shaved nearly a full minute from those times at 185 seconds. In the other direction, copying the data from the drives, Crucial sprinted ahead at 130.7 seconds, but the mechanical Momentus drive wasn't far behind at 144.7 seconds. Ridata and the Barracuda were third and fourth at 156.8 and 166 seconds, respectively.”

More details here.

FaceBook server infrastructure

2008-04-27T09:32:00.003+05:30

Om Malik reported on Gigaom.com:

The company is running around 10,000 servers, according to Data Center Knowledge, citing comments made by Facebook VP of technology, Jeff Rothschild, at a recent MySQL user conference. (See video of the panel.) Of the 10,000 servers, 1,800 are from MySQL and around 805 of them are memcached servers. In order to house its sprawling infrastructure, Facebook has leased data center space from DuPont Fabros in Ashburn, Va., and Digital Realty Trust in Santa Clara, Calif., DCK reports.
How much is Facebook spending on its infrastructure? The company isn’t going to tell us, but there are clues. Server and storage company Rackable today reported first-quarter 2008 sales of around $69 million. Facebook is one of its largest customers, accounting for around 10 percent of Rackable’s sales (that number could be higher, but we’ll have to wait for Rackable’s 10-Q to get a clearer picture), so some quick, back-of-the-envelope math reveals $7 million in spending by the social networking company. A well placed source of mine just let me know that Facebook is going to spend over $9 million more on servers this year. That should be good news for Rackable. Next on my list is an estimate of Facebook’s bandwidth and data center costs.

Doing a little more calculation - $7M for 10,000 servers, means about $700 per server, assuming a 10% cost for F5,firewall etc. - we are at $630 per server. That's pretty heavy duty server, I think.

Komli to represent eBay India for all their ad sales worldwide

2008-04-03T18:28:00.002+05:30

Komli and eBay India have entered into an exclusive partnership whereby Komli will represent eBay India for all their ad sales worldwide. In addition, Komli's ad network optimization technology PubMatic will optimize eBay's unsold ad space for maximization of revenue.

This is very exciting news for a couple of reasons:
1. A global internet giant has chosen to partner with an Indian startup for its superior understanding of online advertising and online advertising technology,
2. This bodes well for the growth of online advertising in India -- large portals, which in the past have not looked at online advertising as a key revenue driver, are starting to do that now.

For details see official news release at - http://www.komli.com/news/ebaypress.php .

Iframe ad-tag vs. Script ad-tag: Online advertising tag type comparison

2008-03-30T12:27:00.008+05:30

This is a list that I have discussed many times with friends, however I never found these on a single place so here you go ...

Differences between iframe tag and script tag:

Iframe tag does not delay the loading of the web-page elements: Iframes usually load in parallel, so for example if you have several elements in a page like images, CSS, JavaScripts and HTML tags and you have the ad-tag as an iframe embedded in the page, the iframe loading would happen in parallel and it would not make your page loading slower. So, if you want page to load faster use iframe tags.
Script tag does not change the “referrer” property of your ad-tag: If your ad-tag is served from inside an iframe, the ad-network that serves the ad will see a referrer property different that your page url/domain. On the other hand if you use a script tag, then the referrer url remains the same as your page url and therefore your domain name. Some ad-networks that require that the ad being served from the same domain that they were created for, will therefore not work with iframe tags (therefore they will not serve ads). Most ad-networks however allow setting of a “site-alias” that allows you to set a different domain from which the ad may be served. Read more about the referrer property here.
Script tag works better for ad-networks that do contextual analysis of the content of the page: if you use iframe tags, ad-networks will not be able to look outside of the iframe therefore they will not be able to do on-the-fly contextual analysis of the contents of the page, therefore they may serve irrelevant ads. Read more about contextual analysis here.
If there is more than one ad from the same ad-network, and you are using iframe tags, these ads may not be able to communicate amongst themselves since the scope of the JavaScript variables is within an iframe. Therefore if an ad-tag sets a JavaScript variable, which the other ad-tag on the same page is expected to read, this will break if you use iframe tags.
Since JavaScript variables have their scope only within that iframe, they don’t contaminate the namespace of the JavaScript variables of your web-page, neither do they get affected by the JavaScript variables of your web-page.
Iframe tags are easier for inclusion inside a web-page, since you can save an ad-tag in a file, and load it as an iframe into your web page. This will also allow parallel load of the ad-tag iframe. For example if your web-page is:

<html>
<script type=”” …>
</script>
<iframe
src=”ad-tag.html”></iframe>
…
…
<body>
</body>
</html>

More questions? Drop me an email.

Update: For #3 "Script tag works better for ad-networks that do contextual analysis", Google AdSense does mention in their help section for Why aren't my ads relevant?, read on:

The AdSense code was placed within an IFRAME.
Our targeting technology is not optimized to serve ads within a separate IFRAME. If you placed the AdSense code in a separate IFRAME, your site may display less targeted ads or public service ads. For better results, please implement our ad code directly into the source of your webpage. Once you make these changes, relevant ads may not appear immediately. Until we are able to re-crawl your site, which may take up to 48 hours or more, your page may continue to display untargeted or public service ads.

Disk storage - where are we headed?

2008-03-20T12:56:00.005+05:30

Some insightful articles and some of my own thoughts on the trends in data storage:

THE BACKGROUND:
Disk capacities are going up and costs are going down, however the effective transfer bandwidth (ETB) per byte of capacity has come down tremendously. Despite capacities and transfer rates increasing by factors or 10,000 and 100 respectively, typical drive ETB has actually decreased by a factor of 100. As Jim Gray said "Disks have become tapes." (Link to source).

Consider, for example, a 10 TB database. Ten years ago, this database would have occupied two thousand 5 GB drives - a common size at the time. With a 3 MB/second transfer rate, the aggregate bandwidth of these 2,000 drives would have been 6 GB/second, enabling the entire database to be scanned in about 30 minutes. Today, only about 20 higher-capacity drives would be needed to hold this same database. Those 20 drives would have an aggregate bandwidth of 1.2 GB/second, increasing the time required to scan the entire database to 150 minutes - an increase of two hours.

DISKS ARE BECOMING A SEQUENTIAL ACCESS DEVICE RATHER THAN A RANDOM ACCESS DEVICE
Jim Gray points out - We have to convert from random disk access to sequential access patterns. Disks will give you 200 accesses per second, so if you read a few kilobytes in each access, you're in the megabyte-per-second realm, and it will take a year to read a 20-terabyte disk. If you go to sequential access of larger chunks of the disk, you will get 500 times more bandwidth—you can read or write the disk in a day. So programmers have to start thinking of the disk as a sequential device rather than a random access device.

Tom White later says that - "MapReduce is a programming model for processing vast amounts of data. One of the reasons that it works so well is because it exploits a sweet spot of modern disk drive technology trends. In essence MapReduce works by repeatedly sorting and merging data that is streamed to and from disk at the transfer rate of the disk. Contrast this to accessing data from a relational database that operates at the seek rate of the disk (seeking is the process of moving the disk's head to a particular place on the disk to read or write data). Read more here.

My take is that SSDs are going to take a while to become an economically viable alternative to disks. Flash disks cost approximately $10/GB, and the OEM costs of good flash drives cost about $60/GB or more (source here). Compare this with the cost of disk, which is about $0.20/GB. So, we are looking at about 300x price difference here. So, I think, it's going to take while before SSDs become reality in storing terabytes of data. Until that time, we will have to use 50-70% empty disks to enhance striping-performance. So, if we were to use 50% empty disks, the cost of disks doubles for storing the same amount of data.

MySQL - Is there a theoretical limit?

2008-03-20T11:30:00.005+05:30

Guy Kawasaki interviewed Marten Mickos. Marten was the CEO of MySQL, now he is the senior vice president of the database group within Sun.

Interesting question about MySQL scalability:

Guy: Is there a theoretical limit of MySQL in case a small business uses it and sales/transactions/whatever explode?
Marten: Every software product has its limits, but I think we have shown that MySQL can scale enormously. Google runs its entire ad system on MySQL. Nokia runs mobile phone networks on MySQL. Booking.com runs all their business transactions on MySQL. If a small business reaches those limits, it is not a small business any longer--it is an enormous global player.

I am a fan of MySQL myself.

Good to know that FaceBook also uses MySQL (with it's 65 million users), and YouTube used MySQL and Twitter uses MySQL (Scaling Twitter: Making Twitter 10000 Percent Faster).

TaffyDB - A JavaScript DB worth trying out

2008-03-19T12:14:00.001+05:30

I recently read about TaffyDB, tried it today. Seems like a handy tool. I would like to use it. TaffyDB is a JavaScript Database, something that can be used for offline data processing in my opinion. For example, a relevant use case is I would like to cache a large report on my browser side and present different views by querying the TaffyDB (I would not like to make server side calls).

It seems like previous attemps have been made for a JavaScript database, a few example are - JavaScript SQL Database with Permanent Storage, Simple JavaScript Database, etc.

TaffyDB is pretty simple to use. Seem feature rich - Under 10K, CRUD Interface (Create, Read, Update, Delete), Sorting, Advanced Queries etc.

Code is pretty easy to write too. Pretty cool, check it out.

Leadership is influence ...

2008-03-15T11:02:00.003+05:30

Stephen Covey writes about Gandhiji's leadership, excellent paragraph on leadership that I read recently:

"People think that leadership is a position. It isn’t. Leadership is influence. The key to influence is what we’re talking about. You can have influence without position. So don’t be so dependent upon position or formal authority, but use your moral authority, what you know is right."

Read more here.

Internet Explorer 8: My Top 10 list

2008-03-10T18:50:00.002+05:30

IE8 has a bunch of really nice features. I read a few reviews (here and here) and made my ‘Top 10’ list, read on …

AJAX Back Navigation enables users to navigate back and forth without leaving the AJAX application and could be used navigating a page without performing a traditional full navigation. This allows websites to trigger an update to browser components like the address bar by setting the window.location.hash value, firing an event to alert components in the page and even creating an entry in the travel log.
DOM Storage is a simple-to-use method for storing and retrieving strings of key/value pair data. Data can be stored per tab instance for a session or persisted to the local machine. This allows pages to cache text on the machine which reduces the effect of network latencies by providing faster access to pre-cached data. Several innovative uses are possible. For example, use this in combination with the new network connectivity event to allow a page to cache data if it detects that the computer is offline.
Six connections per host instead of two for broadband scenarios and a scriptable property allow for more improved performance by allowing parallelization of downloads in Internet Explorer 8. In addition, this increases functionality by ensuring a request is not blocked to a host if two connections already exist. Websites can optimize their downloads based on a scriptable property.
WebSlices - WebSlices is a new feature for websites to connect to their users by subscribing to content directly within a webpage. WebSlices behave just like feeds where clients can subscribe to get updates and notify the user of changes. A WebSlice is a portion within a webpage that is treated like a subscribe-able item, just like a feed. To enable a WebSlice on your website, annotate your webpage with class names for the title, description, and other subscribe-able properties.
Offline Events - This is an easy way of detecting connectivity within the confines of JavaScript. With it we can write graceful offline Ajax applications. Firefox 3 and IE 8 appear to be the only browsers to support this feature.
Cross-domain Request (XDR) - XDomainRequest, is the easiest way to make anonymous requests to third-party sites that support XDR and opt in to making their data available across domains.
Cross-document Messaging (XDM) APIs allow communication between documents from different domains through IFrames in a way that is easy, secure and standardized.
Facebook Integration: Microsoft capitalized on their partnership with the popular social networking site, Facebook, to allow IE8 users the ability to get status updates from Facebook right from their browser toolbar.
eBay Integration: Like Facebook, this feature also uses IE8's new technology, called "WebSlices", which introduces a new way to get updates from other sites via the browser itself, without having to visit the web site.
Firebug for Internet Explorer - We finally have a heavily-Firebug-inspired tool inside Internet Explorer. To quote Joe Hewitt (creator of Firebug): "I couldn't be happier that Microsoft completely copied Firebug for IE8." I have to agree - a tool like this has been a long time coming and it's greatly appreciated. Only the Internet Explorer team would've ever been the ones to build this tool - there's simply too much information here that's unavailable to typical IE extensions.
Browser mode toggling - At first glance this feature makes the most sense for seeing if your IE 7 page will work ok in IE 8. In actuality, however, this will end up being very useful for developing a standards-compliant page (in IE 8, FF, Safari, Opera) and then toggling to see what the result is like in IE 7. This is so much better than the IE 6 to IE 7 jump where you have to keep your browser in a virtual machine in order for it to run side-by-side (according to Microsoft, at least - even though there were standalone solutions).

Read more at the Microsoft IE8 readiness site.

CAPTCHA is Dead, Long Live CAPTCHA!

2008-03-09T14:18:00.002+05:30

Interesting post on coding horror. 3 of the most well known CAPTCHA's are now broken - Google, Hotmail and Yahoo!

Wisdom comes from Gunter Ollman, he notes:

CAPTCHAs were a good idea, but frankly, in today's profit-motivated attack environment they have largely become irrelevant as a protection technology. Yes, the CAPTCHAs can be made stronger, but they are already too advanced for a large percentage of Internet users. Personally, I don't think it’s really worth strengthening the algorithms used to create more complex CAPTCHAs – instead, just deploy them as a small "speed-bump" to stop the script-kiddies and their unsophisticated automated attack tools. CAPTCHAs aren't the right tool for stopping today's commercially minded attackers.

Read more here.

Skype All-Hands: Works really well

2008-02-20T13:33:00.002+05:30

I did a "Skype All Hands" this morning. Surprisingly it worked much better than a face to face all-hands, or a teleconference-all-hands. To be specific the problem that I mostly face is - people don't talk, they don't ask questions during such all-hands meetings. In a face to face all-hands meeting, it takes a while before the first guy asks a question, and then the second guy, and many questions come towards the end of the meeting. A Skype all-hands on the other had turned out to very interactive, people asked many questions, they really participated in the meeting. It seems like engineers like typing much more than talking. Well, I love this. As an added advantage - you already have the meeting minutes (cut-and-paste from IM log), and you can do this across the oceans.

MOSSO is good - but where is my SSH and how much memory do you support?

2008-02-19T20:13:00.002+05:30

TechCrunch reported - "Hosting provider Rackspace is offering a new cloud computing service through its subsidiary Mosso. The service competes with Amazon’s Elastic Compute Cloud (EC2), although it doesn’t require any load balancing or other administration. It also competes with Joyent and Media Temple’s Grid Service. Pricing starts at $100 a month for - 50 GB of storage, 500 GB of bandwidth for transferring data and 3 million HTTP requests. From there additional capacity per month costs: $0.50/GB of storage, $0.25/GB of bandwidth and $0.10/1,000 HTTP requests."

All this is good, but where is my ssh? Dude, how will I install my custom built software? How will I manage my Apache expire headers, how will I implement my mod_rewrite rules?

Also, it's not clear how much memory does the $100 get me?

3 Steps to Adrenaline High

2008-02-07T11:04:00.000+05:30

Here are my 3 steps of getting ‘Adrenaline High’ at 8AM; what are yours?

Pump Iron: 44lb on wrist curls, 105lb on pec-dec, gives me a high that no other drink can

Music: Linkin Park at 88 decibels, so I can’t hear anything else – gives me a high at 9:15AM

Keeping schedule to the minute: My morning schedule runs at a granularity of 1 minute, between 7:55 and 8:09AM there must be at lease a dozen things getting done. Getting things done ‘right’ at the 1-minute-granularity gives me a high (I think it has a name, it’s called ‘urgency syndrome’)

PubMatic Enables Ad Optimization Across Every Ad Network

2008-01-29T11:28:00.000+05:30

Palo Alto, Calif. - (January 28, 2008) - PubMatic (www.pubmatic.com), the first and largest ad optimization platform for Web publishers worldwide, today announced the ability to optimize online ads across any and every ad network. Now Web publishers using PubMatic can eliminate the headache of testing and deciding which ad network and layout will maximize their revenues, because PubMatic does it for them.

Currently in beta, PubMatic serves more than 2,000 publishers and more ad networks than any other ad inventory optimization platform.

"PubMatic immediately doubled our ad revenues by recommending the optimal ad network for each and every visit to WinCustomize.com," said Michael Crassweller, Web Site Manager, StarDock. "Since Wincustomize.com serves up nearly 4 million ads per day, PubMatic's ad network optimization has made a big difference to our bottom line."

The PubMatic public beta is open to all Web publishers, regardless of geography or company size. Signing up is simple and free: publishers can visit www.pubmatic.com/signup to get started in minutes.

5 Attributes of Highly Effective Programmers

2008-01-26T09:28:00.000+05:30

Very nice article on Top 5 Attributes of Highly Effective Programmers:

What attributes can contribute to a highly successful software developer versus the ordinary run-of-the-mill kind?
Humility
Once you start assuming you’re the expert and final word on something, you’ve stopped growing, stopped learning, and stopped progressing. Pride can make you obsolete faster than you can say “Java”.
"The competent programmer is fully aware of the strictly limited size of his own skull; therefore he approaches the programming task in full humility, and among other things he avoids clever tricks like the plague." - Dijkstra
Love of Learning
Good judgement comes from experience, and experience comes from bad judgement. - Fred BrooksObviously, some balance has to be struck here. You cannot learn everything–it simply isn’t possible. Our profession is becoming increasingly specialized because there is simply too much out there. I also think that in some respects, you need to love learning just for the sake of learning.
Detail-orientedness
The steps of changing a piece of software could be:

Thoroughly understand what the code is doing and why
Look for any and all dependencies and interactions with this code
Have a well-thought-out mental picture of how it fits together.
Examine the consequences of changing the feature.
Update all related code that needs to (and repeat this cycle for those components)
Update auxiliary pieces that might depend on this code (build system, installer, tests, documentation, etc.)
Test and repeat.

Adaptability
For me, the first step in changing my mind set is to not get frustrated every time things change (”But you specifically said we were NOT going to implement the feature to work this way!”).
Passion
Ok, that’s maybe a bit of exaggeration, but by listing the counterpoints, it’s easier to see symptoms of someone who does have passion:

Thinks and breaths technology
Reads blogs about programming
Reads books about programming
Writes a blog about programming
Has personal projects
These personal projects are more important than the boring stuff at work
Keeps up with latest technologies for their interests
Pushes for implementation of the latest technologies (not blindly, of course)
Goes deep in technical problems.
Not content with merely coding to spec.
Needs an outlet of creativity, whether it be professional (software design) or personal (music, model building, LEGO building, art, etc.)
Thinks of the world in terms of Star Trek

Just kidding on the last one…
…(maybe)

Read full article here.

8 hacks for finding Startup office space

2008-01-18T18:12:00.000+05:30

I started looking for a new office for Komli Engineering at Pune, India about a week back. Here I describe my journey and the final selection.

Here are some of the key points when finding an office of about 2000 sqft in Pune in Aundh/Baner area:

Rates have gone up like crazy – average rate is Rs. 50/sqft., unfurnished
Most office spaces have only 2 restrooms, which is too few for a 2000 sqft space. So most spaces can pretty much be rejected on that ground
There are a large number of residential properties that people are converting into commercial properties for offices and showrooms, and charging Rs. 50 per-sqft!
The problem with these residential-turned-commercial properties is that – a) families are living in the same building, b) kids are playing in open spaces and c) parking is mostly an issue.
There are independent-bungalows available at very cheap rates. These places are great – they are peaceful, have lots of spaces, lots of parking and so on. BUT you would probably never get broadband in those places. These independent-bungalows are available at 1/3 the rental cost
Your office space must be located not more than 3 minutes from 5 places that must sell wada-pav, hot samosas, cut-chai, tandoori chicken and Pizza Hut – else you are doomed, because most employees in a startup are not married, and they need to eat (when they are not writing code)
The other most important things when you are renting an office space are – 1) the place is good, 2) broadband is feasible and 3) parking space is available. The “broadband” is the most unexpected thing to find out. You can find the best place and the least cost, but no broadband – that will totally blow you off. The second most difficult thing to find is parking space for 4 cars
I looked at the most cool places such as a nice place next to McDonalds in Aundh, a really cool office with all glass façade – but not good enough for Komli!

I finally decided with a really nice place above “Kobe Sizzlers” in Aundh. Awesome place, lots of space, central location, 2 balconies and lots of eateries around. And the best part is – you can get Sizzlers on Demand.

Wanna join us - check-out our open positions at http://www.komli.com/careers/ .

How to recognise a good programmer?

2008-01-12T12:08:00.000+05:30

Daniel has written a really great article about how to recognize good programmers. Here is a summary of the traits (I have marked the ones that I truly relate with, and added my comments):

Positive indicators:

Passionate about technology
Programs as a hobby
Will talk your ear off on a technical subject if encouraged [Mukul: This is an absolute necessity to identify a "hacker". They are so passionate about what they do that they will talk about it for more time than you want them to. I love this feature.]
Significant (and often numerous) personal side-projects over the years
Learns new technologies on his/her own
Opinionated about which technologies are better for various usages [Mukul: I have seen some of the technical guys getting upset when you tell them do use a certain technology or tell them to do things in a certain way - this is good. They should be opinionated.]
Very uncomfortable about the idea of working with a technology he doesn’t believe to be “right”
Clearly smart, can have great conversations on a variety of topics
Started programming long before university/work
Has some hidden “icebergs”, large personal projects under the CV radar
Knowledge of a large variety of unrelated technologies (may not be on CV)

Negative indicators:

Programming is a day job
Don’t really want to “talk shop”, even when encouraged to
Learns new technologies in company-sponsored courses
Happy to work with whatever technology you’ve picked, “all technologies are good”
Doesn’t seem too smart
Started programming at university
All programming experience is on the CV
Focused mainly on one or two technology stacks (e.g. everything to do with developing a java application), with no experience outside of it

Read the full article here.

Thrudb: Better Storage?

2007-12-30T17:57:00.000+05:30

I recently read about thrudb, and I must say I am very impressed with the lucidity with which Jake Luciani describes the problem and the solution. Here is an excerpt:

"Data on the web is often fluid and loosely structured and it is becoming increasingly difficult to fit this data into a fixed database schema which is amended over time. A simple example of this is tagging. The many-to-many relationship of tags is difficult to query efficiently using tables and SQL, such that ad-hoc solutions are required.
Also, web data is often "mashed up" and viewed together (e.g. Facebook profile) or viewed spatially (e.g. Google maps + event data).
In order to provide this new kind of data flexibility the web is moving towards a document-oriented data model, where records aren’t grouped by their structure but by their attributes.
There are also standard data-oriented issues like indexing, caching, replication and backups, which are left for "later" but are never easy to implement when it’s time to do it. There are a number of great of open source solutions to these problems, but they require proper integration and configuration. These components end up being learned over time and learned by trial and error.
Thrudb, therefore, is an attempt to simplify the modern web data layer and provide the features and tools most web-developers need. These features can be easily configured or turned off."

Looks very cool. I am going to try this out as soon as I get hold of my developer box tomorrow morning.

Thrudb talks about the following features:
• Client libraries for most languages
• Multi-master replication
• Incremental backups and redo logging
• Multiple storage backends (S3 included)
• Built for horizontal scalability
• Simple and powerful search api (Lucene)

My Laptop Search: From Lenovo to DELL

2007-12-19T12:57:00.000+05:30

My laptop search started 2 weeks ago. I had some very simple requirements; however I found it incredibly tough to meet my simple requirements. I am specifically talking about buying a laptop in India, specifically in Pune.

My simple requirements were – laptop should:

Have an Intel Core 2 Duo based mother-board
Should have at least a 667MHz FSB
Should have at least 2MB L2 cache
Should have a 14” screen
Should be light weight
I assumed that all laptops would have at least 80GB HDD and 1GB RAM

So, my search began with Lenovo, I looked at the X and Y series. Surprisingly, Lenovo, uses some really old CPU, most of their mother-boards were based on 533MHz FSB, the Intel Core 2 Duo mother-boards with 667MHz FSB cost more than Rs. 45K, which is a lot. I looked around at all available models (in India), couldn’t find a model that would cost me less than Rs. 50K (Lenovo 3000 Y300/Y410/Y500).

Then I looked at Toshiba. They had some really good mother-boards, for example the Satellite M200-A411D. However the look and feel of the laptop looked more like my 5 year old son’s laptop. There was a lack of ruggedness, lack of proper ergonomic thinking for the keyboard and port design. Specifically the keyboard is white in color, which means that it will get dirty in a week, and I won’t be able to fix that.

I looked at Sony Vaio VGN-CR23G series. Really cool models. Nice colors, nice mother-board, nice keyboard, almost everything was nice. Except for the weight – it weighs 2.5 kilos, which is more than I thought it should be. The cost was also higher, at Rs. 54,990 .

Then I saw DELL VOSTRO 1400 Notebook. I got an Intel Core 2 Duo Processor (T7250), 2.0GHz, 2MB Cache, 800 MHz FSB, 1GB RAM, 120GB HDD, 9 cell battery, and Vista home. That was the end of my search. I found my laptop. I bought it locally instead of ordering it online or on phone. Turns out to be a pretty good laptop, I happy with it. DELL has got much better since I started using DELLs back when I was at VERITAS in 1999-2002 timeframe. Good job DELL.

Looking back, Lenovo is way behind in its CPUs, I think. I love the rugged design of Lenovo, but they should get newer CPUs faster FSB and more L2 cache.

My dream machine - Nikon D40X DSLR

2007-11-27T18:36:00.000+05:30

My Dream Machine, finally:

I purchased this from jjmehta.com on 25th November 2007.

Camera information display and settings

Some specialties about D40X:

The most compact Nikon digital SLR ever, featuring intuitive controls and ergonomics designed for everyone
10.2-effective-megapixel
Fast continuous shooting mode enables up to 100 JPEG images (FINE L-size or smaller) at 3 frames per second
High-precision digital image processing algorithms for natural-looking pictures with faithful color and tone reproduction
Automatic and manual control over ISO sensitivity from ISO 100 to 1600, as well as HI 1 (manual only)
Quick 0.18 second power-up to respond to every photographic opportunity
A large 2.5-inch LCD monitor with viewing angle of approx. 170 degrees in all directions
Long-life rechargeable lithium-ion battery that allows up to 520 images* on a single charge (*CIPA standards)

algoGod update: Extending the submission date to December 14th 2007

2007-11-14T15:22:00.000+05:30

November 14 2007: Here is an update on the algoGod contest.

First, we are extending the submission date to December 14th 2007.

We started the contest 4 weeks ago, and got an overwhelming response with 299 contestants registering for the event and actively solving the problem. Many contestants requested that the submission date be extended, so we have extended the last date.

Also, we would like to inform you that we are hiring machine learning experts who can work at Komli for making a world-class ad-optimization engine. If you are interested – check out our web page at http://www.komli.com/careers under ‘Machine Learning Expert’. We introduced a FAQ for algoGod, it is on the following web page: http://www.komli.com/algogod/faq.php If you do have any questions, please don’t hesitate sending it to algogod@komli.com.

AOL buying Quigo

2007-11-07T22:08:00.000+05:30

AOL is buying Quigo, a contextual ad network, for a reported $340 million [techcrunch.com]
Quigo will be the fourth ad company AOL has acquired in 2007. Earlier in the year, AOL acquired Third Screen Media (mobile advertising), Adtech AG, an ad serving platform based in Frankfurt, Germany, and Tacoda, the behavioral targeting company. All of them roll into Platform-A, AOL’s recently announced ad platform division [paidcontent.org]
Quigo, has over 500 publisher relationships and about 3,000 advertisers. [readwriteweb.com]
Quigo has raised $45 million since opening its doors in 2000 [paidcontent.org]

algoGod: Be Crowned the World's Greatest Algorithms Expert!

2007-10-16T17:35:00.001+05:30

Komli today launched the “algoGod contest” for machine learning, math, genetics, and algorithm experts.

http://www.komli.com/algogod/

Contest winner to receive Rs. 2,00,000

Start date: October 15th 2007.
Entries must be submitted on or before November 14th 2007.
Results will be declared on December 31st 2007.

Have you ever wondered if you are the best algorithms expert on the planet? Have you ever thought, "I know I can beat everyone, just let me prove it?" Well Komli's algoGod contest is for you, it's your chance to show the world how smart you really are!

The algoGod contest seeks to crown one expert as the 'Algorithms God'. How are we going to do this? Well, the proof is in the pudding! We want every contestant to solve a common problem, and whoever is best will receive the algoGod prize!

A little more background:

Komli lives in the world of online advertising, and online advertising is rife with opportunity with complex algorithms based on cutting edge topics such as machine learning, data mining, graph theory, etc. Online advertising is growing at a very fast pace, and the number of variables affecting the performance of an online ad has been growing at an even faster pace. Komli is devising methods for maximizing the yield of online advertising using advanced statistical machine learning methods over large-scale systems. This is a very interesting and complex algorithm problem.

Komli is currently using a set of algorithms for maximizing the yield of online ads, collectively called 'Yin-Yang'. There are a lot of interesting alternative approaches to Yin-Yang that have yet to be tried. Komli is interested in determining if any of these alternative approaches can beat Yin-Yang by making better predictions.

Komli will provide participants with anonymous ad impression data and a prediction accuracy bar that is 50% better than what Yin-Yang can do on the same training data set. Participants' solutions will be judged by 'Time complexity' and 'Space complexity' criteria. The participant whose solution works best will receive Rs. 2,00,000, bragging rights and an opportunity to work with Komli. Of course, participants have to share their method and code with Komli. Eager participants can signup for the contest by filling the form on the left. Also, please let us know of any questions at algogod@komli.com.

PubMatic Engineering: Slide show

2007-10-04T22:24:00.000+05:30

PubMatic selected by TechCrunch as a Top 40 Startup in the World

2007-09-30T19:49:00.000+05:30

PubMatic, a product of Komli, was selected by TechCrunch as a Top 40 Startup in the World. Nearly 750 startups from around the world applied for this honor, and PubMatic was lucky enough to be selected! This was announced at the TechCrunch40 conference in San Francisco, CA, a conference built to showcase these 40 top startups.

We are hiring!

If you dream in Java, think in PHP, and talk in <xml> over IM, you should talk to us.

In addition, as part of our presentation at the conference, we announced that PubMatic has been released into a global beta available for all publishers around the world! During our alpha over 500 publishers from around the world have been using PubMatic and seeing some amazing results. See news about PubMatic here.

Online advertising is growing at a very fast pace, and the number of variables affecting the performance of an online ad has been growing at an even faster pace. Komli is devising methods for maximizing the yield of online advertising using advanced algorithms running over large-scale systems. We are also developing decision support system for data analytics, analysis of real time data, such as user behavior and web analytics, server scalability to support 100,000,000 requests per day (to start with), and much more cool stuff.

The last I posted about Komli, we had just moved into our new office. We were still building the product. Since then a lot has changed, we wrote a bunch of code, did a beta, were selected as a Top 40 startup in the world, our team grew to 8 people, and have been having a lot of fun.

The beta release was amazing, we had close to 400 customers using PubMatic, a small team of very enthusiastic world-class programmers were writing code and managing escalations at the same time.

While we hacked code in Java, PHP, AJAX and C 12 hours a day, and listened to rock and the latest Bollywood tunes of Bhool Bhuliyaa, the continued to have a sense of humor. This is a sketch that one of us drew on the whiteboard, while he was designing a new DB schema for user authentication.

And, did I mention, we never miss a chance to have fun ...

Simple Tooltip

2007-09-30T18:28:00.000+05:30

http://codeeazy.com/mktooltip/

I wanted a tooltip implementation for my web page, and was looking around. I looked a number of open-source options (see 40 tooltip scripts here and 20 here) and one commercial library. I liked the functionality however each of them was 7, 8, 10 or 25 KB in size, which I thought was too much to implement just a simple tooltip. So, I thought how about I write a tooltip library myself.

Here is an implementation, check it out at http://codeeazy.com/mktooltip/. I implemented this in 818 bytes of code, and about an hour of coding. So one would think, there must be something wrong. Well, it works. And is cool!

Three kinds of platforms you meet on the Internet - in Mike's words

2007-09-18T15:33:00.000+05:30

Please read this insanely useful article written by Mike Anderson.

Mike has described 3 kinds of platforms for web services.

Level 3 platform is the most difficult to build by the platform-provider and is very easy to integrate and use from an end-user’s-perspective. While, Level 1 platform is easiest to build by the platform-provider and difficult to integrate from an end-user’s-perspective. Most people build level-1 platforms – like whenever you see the work “API” it actually means a Level 1 platform. Exception is Facebook and Ning, which are building a Level-2 and Level-3 API respectively. Level 3 is really an API nirvana, and requires very good architectural planning from the platform-provider side. I can imagine some API evolving from a Level 1 to a Level 3; but there aren’t any examples of such APIs; also the code written for Level 1 API, will probably be difficult to use in Level 3, unless there is again some very good architectural planning.

A must read for anybody who codes!

Read on …

Startup: Focus on the problem

2007-07-27T11:22:00.000+05:30

Peter Rip at Crosslink has posted a great article about how Teqlo failed and now they are fixing it. I am reproducing a few excerpts, and my comments:

First, let me admit we went down a mashup rat hole. We have a general technology for snapping together web services. "Because they can" is an insufficient answer to "why do people want to create mashups?" We failed to commit to solve a specific problem for a specific market, preferring instead the broad appeal of generality. This has changed.

I have seen so many people having great ideas, looking for problems to solve. So basically, you have a solution, and you are looking for a problem. This almost never works this way, even though it seems (obvious enough) that it should work.

The first thing we did was toss out any pretense of solving everyone’s problem. There is an old proverb that I just invented for this situation -- “The boiling of the ocean begins with a single puddle.” We had to define our puddle. So we did.

Peter has hit the nail on the head. Trying to solve all problems at once is one of the biggest problems you can face. You need to have laser sharp focus on one specific problem area that you want to solve. It is very possible that 80% of your friends will tell you “why don’t you also do this, this is in adjacent space (technologically)”, “if you don’t do this, you will miss out on all the opportunities in that (other space)”. My advice is to tell your friends is “we will get to that”.

Excel tip: Instant in-cell graphs

2007-07-26T22:24:00.000+05:30

Lifehacker has posted a Really cool hack:

Ad Networks: Why it’s better than ever to be a targeted content site

2007-07-25T11:14:00.000+05:30

Jeremy Liew has written a very interesting article on Venturebeat:

A recent report by Collective Media found that:

66% of advertisers plan to increase their usage of ad networks in 2007
88% of respondents planning to use online ad networks in 2007 (up from 77% in 2006).
57% of respondents believed how an ad network targets audiences was the #1 differentiating factor between networks

...

The WSJ recently had an excellent article on behavioral targeting that detailed Pepsi’s launch of Aquafina Alive,their new low cal vitamin enhanced water. The campaign was backed by an online campaign through Tacoda and targeted to people who had previously visited “healthy lifestyles” websites.

The result? Pepsi recorded a threefold increase in the number of people clicking on its Aquafina Alive ads compared with previous campaigns. “We’ve never been able to get to this level of granularity,” says John Vail, director of the interactive marketing group at Pepsi-Cola North America.

Mobile Bliss

2007-07-21T20:20:00.000+05:30

I was reading a post by Fred Wilson on AVC.

From today's front page story in the New York Times about Google's $4.6bn wireless bid.
In the Internet giant’s view of the future, consumers would buy a wireless phone at a store, but instead of being forced to use a specific carrier, they would be free to pick any carrier they wanted. Instead of wireless carriers choosing what software goes on their phones, users would be free to put any software they want on them.
Hell yeah! This is the way it must be. Open devices, open services, open spectrum.

What would be really cool is if Google paid $4.6bn for the spectrum and then opened it up for the world to use as we see fit, just like Facebook opened up their platform.

The good part is that we already have this in India. It's strange that we don't realize the importance of things that come free to us - I can go any buy a phone, choose whatever carrier I want, and put whatever software on that, TODAY!

CodeEazy - solving problems, Simply! - Launching my other blog

2007-07-20T16:02:00.000+05:30

Anand and I have been talking about this blog for a while, there was just no time to write about some of our inner feelings about - coding. So today we launched our other blog - CodeEazy.

We have been working on, what many people would call “the most bleeding edge” technology at Komli. To give a little bit of background, we are the founding engineers at Komli. Komli is an early stage startup. Since last November we have been learning some very cool techniques of solving problems. When it comes to building a product, a typical engineering mindset, lots of times gets stuck in technical jargons, technical-coolness and sometimes religious-warfare between different technical solutions, languages etc. While, the most important thing when you are building a product is – to solve a problem. That’s it. You have to solve a problem, apply any technology, apply any language, apply any tool, use two languages or 3 languages – you gotta solve the problem. While, I obviously, am against making code unreadable or un-maintainable, and therefore mixing technologies or languages may be a no-no in some cases, however engineers mostly overdo that, by applying that rule too strictly, therefore loosing focus on – solving the problem; you gotta solve the problem, that’s the most important thing.

We have been solving numerous problems, in a number of areas, and sometimes in very cool ways.

We would use this blog to tell you about some of the cool ways of solving complex problems in simple ways. Some of what we have learnt. Some hard-learnt-lessons. I hope that you will like it.

Venturing into venture capital? Be ready.

2007-07-13T13:19:00.000+05:30

Amar Goel posted a great writeup on venture capital. It is a must read for anybody who wants to raise money for their business:

The 3 lines I liked the most are:

“Raise VC money only … If you would like to build at least a $20M business over 3-5 years and ideally a $100M+ business over 5-7 years you are a prime candidate for venture capital. If you want to build a $3M business over the next 5 years do not raise venture capital.”
“It takes a lot of time to raise venture capital. [At Chipshot.com] Turning that interest into funding took 4 months of almost full-time work. I basically spent 15 hours a day on phone, email and in meetings selling the heck out of my company and our team. ”
“Most VCs will be loathe to invest in some “idea” you have that will change the world that nobody has ever done before.”

I have met many entrepreneur-aspirants who want to raise VC money, but have very little idea of how soon they can get paying customers, and how much those customer are going to pay them and why. I believe it is very very very important to have a clear idea of the business - who is the customer, why will he use your software (or hardware), why would he pay you for the product, how much will he pay you for the product, how would you approach this customer, how often will the customer pay you, how sticky is the product (I mean will they pay once and that's it), etc.

Internet Ad Spends In India To Double In 2007

2007-07-10T19:08:00.000+05:30

Reported on ContentSutra by Anupama Chandrasekaran:

Media agency ZenithOptimedia expects Internet ad spend in India to more than double in 2007 and to be 10 times its current size in 2009. [via agencyfaqs] Ad spends on the Internet could jump to Rs. 450 crore this year from Rs. 210 crore in 2006 and could skyrocket to Rs 2,250 crore mark by 2009, the study said. This would mean that in two years, ad spends on the Internet will surpass those on radio, cinema and outdoor, individually.

Mukul Kumar's Blog

Great startup advice from Ben Horowitz

More depth with smaller ISO in DSLR photography

Google getting hypersocial

PubMatic Pune Mixer on 26th November

Does IT outsourcing industry fear cloud computing?

Concepts 2010 at PICT Pune

The Return of the Salary Hunters

Real Engineers don't give ETAs

Firefox 3.6 has some interesting featureset

Web Scalability & Performance: Real Life Lessons

8 cores and above - Is the race worth it?

Go for that “Impulse Purchase”

Analysis: Which URL Shortening Service Should You Use?

Mikz - How different is it from httpd4mobile and mymobilesite?

Microsoft Planning Ad-Supported Model For Office 14?

Contribute to a Map-Reduce job by simply pointing your browser to a URL

10 Interesting Articles for the Weekend Read

Y Combinator’s FathomDB looks pretty neat

Carol Bartz on 'getting the house in order'

Microsoft to use Machine Learning software to put servers to sleep when not in use

'AWS Public Data sets' has full Wikipedia available in TSV format

Google Apps Status Dashboard looks pretty good

Nokia to enter laptop market?

CDNetworks Acquires Panther Express To Speed Expansion In The U.S.

PubMatic Launches Open, Real-time Monetization API for Ad Networks and Publishers

Google Announces Pricing for App Engine

Safari 4 benchmarked: 42x faster than IE 7, 3.5x faster than Firefox 3

20 million Chinese sites are served by QZHTTP not Apache

Just Start Pitching: The infallible Sales Pitch

Contextually Relevant Ads from Google AdSense that made me click

Some fun with Google Hindi translation!

Top 15 Posts last week, which are still fresh

Yahoo! Email Ads still can’t get geo-targeting right

Komli Launches ViziSense: India’s First Free and Open Online Audience Measurement Platform

5 tips for startups for handling the power situation

The Real Cost of Amazon CloudFront

The world’s most super-designed data center

VMware Wants to Bring Virtualization to Smart Phones - BUT where are the use-cases

Cloud Computing: How important is "data locality" from a costing perspective?

Finding and fixing memory issues using Valgrind with example of Apache+FastCGI web application

Is ‘Divide and Conquer Algorithm’ a human instinct?

5 Hacks for Startup Hiring in India

1. “Offer acceptance is not equal to JOINING”

2. “The Resume”

3. “The Sourcing”

4. “The Interview”

5. “The Timing”

Why doesn't Google Chrome render my bullets right?

EOS 5D Mark II Digital SLR camera - Is this the new future proof SLR Camera?

PubMatic interview with Amar Goel

Chrome shine for Google

Google Chrome Tested

Social network popularity around the world

Servers: What do they really cost?

Cloud Availability

Comparing Clouds: Amazon EC2, Google, AppNexus, and GoGrid

Scientists found a workable way of reducing CO2 levels

Early YouTube Engineer talks about outages, system scalability

Best Web 2.0 Sites in India

Google Advanced Search - now with dates!

Designing Caches for Highly Scalable Web 2.0 Applications

Facebook Photo Storage Architecture

All models are wrong, and increasingly you can succeed without them.

3 awesome resources for Cross Browser Coding

Selling a Used Car in India - eBay.in, CarWale.com, Kijiji.in compared

Web Service for Content Distribution Network

Gartner states the obvious, yet again

World's cheapest airbags at $10 a piece

Twitter and TinyURL

Storage shipments - 1 Billion Gigabytes a year!

How to - Convert Word Docs to Web Pages

TinyMCE

Textism

Performance Showdown - SSDs vs. HDDs

FaceBook server infrastructure

Komli to represent eBay India for all their ad sales worldwide

Iframe ad-tag vs. Script ad-tag: Online advertising tag type comparison

Disk storage - where are we headed?

MySQL - Is there a theoretical limit?