goolge's web page

[xx]

Hi,

Anyone know what technique goolge.com uses for its web page? servlet?
jsp?, or others?

Thanks,

xx

 

[Mark Haase]

Hi,

Anyone know what technique goolge.com uses for its web page? servlet?
jsp?, or others?

Hmm??

If you type something in the search box, it returns a URL like this:

http://www.google.com/search?q=how+does+google+work?&ie=UTF-8&oe=UTF-8

So it's just using HTTP GET, with some propietary software processing
those GET requests.

Unless I've misunderstood you.

PS What's this got to do with Java?

 

[Sudsy]

Hmm??

If you type something in the search box, it returns a URL like this:

http://www.google.com/search?q=how+does+google+work?&ie=UTF-8&oe=UTF-8

So it's just using HTTP GET, with some propietary software processing
those GET requests.

Unless I've misunderstood you.

PS What's this got to do with Java?

It's an interesting question, and Java might be part of the actual
solution. But the underlying mechanism is hidden, which is fair
enough. For all we know they could be using PHP or CGI.
Perhaps the OP is curious as to how to implement a large-scale
platform and the technologies used by one of the leading search
engines.
I can use JSPs and hide that fact from the users of a site. It
doesn't matter to them how I choose to implement my solution.
But from a scalability prespective, people might be curious as to
how I can support a high volume of requests.
We all know that it typically involves a number of different
technologies, ranging from the front-end load balancing to the
maintenance of session information in a back-end database with
exceptional performance.
But asking "how does Google work?" is like asking for trade
secrets...

 

[Jack Jia]

I guess they use Perl. Nothing is good like Perl to do text processing such
as scrawl the Internet.
If they haven't used mod_perl, that is a bit strange.

Jack

 

[Chris Uppal]

But from a scalability prespective, people might be curious as to
how I can support a high volume of requests.
We all know that it typically involves a number of different
technologies, ranging from the front-end load balancing to the
maintenance of session information in a back-end database with
exceptional performance.

Do you know of any good, comprehensive, guide (book or web) to the technologies
that are used and the tradeoffs between them ?

I've been hovering near this area for years, and while I've picked up a certain
amount, I'm sure that I have a very incomplete picture of the available
options.

-- chris

 

[Chris Uppal]

Perl to do text processing
such as scrawl the Internet.

"scrawl the Internet" -- that's an excellent description of Perl ;-)

-- chris

 

[Sudsy]

Do you know of any good, comprehensive, guide (book or web) to the technologies
that are used and the tradeoffs between them ?

I've been hovering near this area for years, and while I've picked up a certain
amount, I'm sure that I have a very incomplete picture of the available
options.

I don't know of any such resource. Technical information tends to be
found in "verticals", with little cross-over. That's somewhat under-
standable since there's a huge amount of knowledge to digest.
One can make a career out of being a router guru, for example. But
then you need to know all about BGP, OSPF and all the other details...
So most router techies might not know about dual-ended SCSI and high-
availability RDBMS servers...or HACMP...or Apache/Tomcat...or...
That's why companies engage the services of consultants like me.

 

[Princess Morgiah]

Hi,

Anyone know what technique goolge.com uses for its web page? servlet?
jsp?, or others?

A quick Google on "how does Google work" gave me the following result:
http://www.cse.ogi.edu/~jsnow/google/google.html

More data and technical facts - everything you always wanted to know
http://www.google.com/technology/pigeonrank.html

Princess Morgiah

 

[Mladen Adamovic]

I don't think so.
I think they use CGI, because for speed (CGI is up to 10x faster then PERL).
They probably use some distribute mechanism for incoming request, and use
the fastest searching algorithm
known so far.

At undergraduate studies usually students learn KMP (Knutt - Morris - Pratt)
searching algorithm, which is not the fastest, but well known and not too
complicated one.

One other well known searching algorithm is Boyce-Codd algorithm, but there
exists a lot efficient algorithms - for those google O;-).

 

[PiedmontBiz]

I don't think so.

I think they use CGI, because for speed (CGI is up to 10x faster then PERL).
They probably use some distribute mechanism for incoming request, and use
the fastest searching algorithm
known so far.

At undergraduate studies usually students learn KMP (Knutt - Morris - Pratt)
searching algorithm, which is not the fastest, but well known and not too
complicated one.

One other well known searching algorithm is Boyce-Codd algorithm, but there
exists a lot efficient algorithms - for those google O;-).

They use Python.
http://pbf.strakt.com/success



allen

 

[Liz]

Google has an API that you can access
after (free) registration. It uses SOAP.

 

[Tony Morris]

I don't think so.
I think they use CGI, because for speed (CGI is up to 10x faster then PERL).
They probably use some distribute mechanism for incoming request, and use
the fastest searching algorithm
known so far.

At undergraduate studies usually students learn KMP (Knutt - Morris - Pratt)
searching algorithm, which is not the fastest, but well known and not too
complicated one.

One other well known searching algorithm is Boyce-Codd algorithm, but there
exists a lot efficient algorithms - for those google O;-).

CGI is to PERL as HTTP is to Internet Explorer.

CGI is a protocol, not a language.
PERL, among other languages (including Java), make use of this protocol for
retrieval and submission of data to a web server (which would otherwise be
stateless and static).

--
Tony Morris
(BInfTech, Cert 3 I.T.)
Software Engineer
(2003 VTR1000F)
Sun Certified Programmer for the Java 2 Platform (1.4)
Sun Certified Developer for the Java 2 Platform

 

[Marco Schmidt]

Mladen Adamovic:

[...]

At undergraduate studies usually students learn KMP (Knutt - Morris - Pratt)
searching algorithm, which is not the fastest, but well known and not too
complicated one.

One other well known searching algorithm is Boyce-Codd algorithm, but there
exists a lot efficient algorithms - for those google O;-).

They are certainly not using string search algorithms.

Instead they have some sort of inverted index that stores all
documents which contain a given word.

Check out the original paper the Google founders wrote on their
project: "The Anatomy of a Large-Scale Hypertextual Web Search Engine"
<http://www7.scu.edu.au/programme/fullpapers/1921/com1921.htm>. A lot
has changed in the mean time, but the major structure has most likely
remained the same.

Regards,
Marco