Random Musings of WWW

***Single Query Fired on Google Needs 1000 multiprocessors…*** 

Google is normally quite secretive about their search infrastructure but, in a break from tradition, they have revealed that a single search query on Google can consume the processing power of 1000 machines.Google Fellow Jeff Dean, in a keynote talk at WSDM 2009, shared some numbers about Google’s impressive growth run from 1999 to 2009. According to Dean, while both search queries and processing power have gone up by a factor of 1000, latency has gone down from around 1000ms to 200ms. Crawler updates now take minutes compared to months in 1999.

Another significant change was the switch to holding the complete search index in memory, resulting in the use of 1000 machines to handle a single query compared to just 12 previously.

This revelation may be a bit embarrassing for Google, which has defended its ecological record in the past, claiming that a single Google query takes just 0.0003KWh of energy and that the Google datacenters are “the world’s most efficient.”

=> Erlang…Computing Concurrently.

Erlang is named after A. K. Erlang. It is sometimes thought that its name is an abbreviation of Ericsson Language, owing to its origin inside Ericsson. According to Bjarne Däcker, who headed the Computer Science Lab at the time, this duality is intentional.

It seems Erlang is a language whose time has finally arrived.
Erlang is poised to be the “Java for the concurrent world”.
Well simply put, it is a functional language which makes concurrentprogramming for multi-core architectures a snap.
Who uses Erlang for product development ? NortelEricsson, T-Mobile, Nokia Research Center and FaceBook.
Recently Facebook launched a Chat Feature that Uses Erlang to Scale up to 70 Million Users.
The reason why Erlang is getting the attention is for following significant developments
The Moor’s law will eventually hit its fundamental limits and the only way to extend it in future will be through multi-core processor architectures. 
There are SMPs in laptops even now and more applications need be build to target that capability.
Most applications today need to process staggering amounts of data using massive parallel systems
Virtually all language except Erlang use shared state concurrency. It is a recipe for disaster as you have to rely on locking mechanisms to allow multi-process/thread access.If the program crashes inside the critical section, recovery is impossible.
Erlang has Message passing concurrency in which there is no shared state and the processes exchange data through asynchronous message passing.
Erlang is used in Disco Project – A Open Source Map-Reduce framework that handles tens of thousands of tasks in parallel.
Erlang seems to be a natural fit to solving high concurrency problems and suited for network server applications that need to be highly distributed yet reliable.
Ok I need to hurry, grab the book “Programming Erlang: Software for a Concurrent World” and start tinkering around.

=> Difference Between “Script” and “Program”:

An amusing quote from Perl speaks it all. However I will explain in a bit more detail here.

“A Script is what you give to the Actor while the Program is what you give to the Audience”.  -PERL.  


=>OpenIDs: –

OpenID is a single username and password given to the user that can be used to login into many websites listed below.



 Orange (France Telecom)

Don’t Have an OpenID Yet?………. 

Many Services that will help you to have  an OpenID.



  • myID.net

    Free OpenID Provider with support for groups and Korean language.

  • myVidoop

    Free OpenID Provider that eliminates passwords with security features, customization, and browser integration.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: