Unfrigginbelievable!!!

[sadie31]

You gotta read this:

http://www.linesave.co.uk/google_search_engine.html

[JPC-Greg]

Pretty impressive but imagine if they take things one step further and turn google into a Seti@hme type project.

[Ron]

Do you mean to distribute the search load across millions of PCs, or to use their machine if and when idle as a participant in SETI?

[the_ancient]

Do you mean to distribute the search load across millions of PCs, or to use their machine if and when idle as a participant in SETI?

Google will not waste time looking for radio signals, Google will just build the damn spaceship and go find them

[Vin DSL]

Heh! To me, this is all predictable, i.e. expected, e.g. not surprising...

The only thing that DOES surprise me is...

Google runs its systems on cheap, no-name IU and 2U servers -- so cheap that Google refers to them as PCs. After all each one has a standard x86 PC processor, standard IDE hard disk, and standard PC reliability -- which means it is expected to fail once in three years.

Then again, why not?

[Spathiphyllum]

Pretty impressive but imagine if they take things one step further and turn google into a Seti@hme type project.

I wonder who would want a Googlescan@home sitting on their PC? Would you trust a company to lurk and archive on your systems?

[JPC-Greg]

Do you mean to distribute the search load across millions of PCs, or to use their machine if and when idle as a participant in SETI?

Idle just like seti used to do and Im referring to using idle pc's worldwide to help archive and search the net, not your own pc... thats an unsettling thought.

[Ron]

I'm guessing they don't have much idle time on their machines, they probably use (or plan to use) all of their non-searching CPU time re-indexing and surfing, and as far as putting search clients on all the millions of PCs around the world I can't think of a way to do that that wouldn't expose their proprietary information to everyone.

Same as searching; the amount of info that would have to be sent across the web is much greater than just reading the page for themselves, and looking at what they were indexing would also give away alot of info about their algos.

[sadie31]

Pretty impressive but imagine if they take things one step further and turn google into a Seti@hme type project.

You mean like this:
http://www.google.com/tools/toolbar/dc/faq_dc.html

[Gwaihir]

This guy has a lot of details that I've not seen public from any recent date. Interesting stuff - if it's based on fact. Anyone know who he is?

As for distributed computing: Google's computers are already quite busy with their "main task" distributed computing, but the Google toolbar that you can install yourself has been computing for Folding@Home for quite some time now. Much more useful stuff IMHO than seti. Folding@Home simulates the folding of specific parts of proteins, parts that have a big impact on the proteins behaviour. Results are already used for medicine and such.

[sadie31]

This guy has a lot of details that I've not seen public from any recent date. Interesting stuff - if it's based on fact. Anyone know who he is?

The guy mentioned in the article is a Google "fellow", he is their VP of Technology if I recall right.

[Connie]

This guy has a lot of details that I've not seen public from any recent date. Interesting stuff - if it's based on fact. Anyone know who he is?

A lot of the information in that article is public information if you read the Patients. The only thing I have not read is in regard to the 30 clusters. But clustering does make sense. One thing that was not mentioned is that results are actually delivered from Data centers and currently there is at leas 70 different data centers. Each data center has a different IP.

Ever done a search, and then repeated the search and get different results? This is the result of search results coming from different data centers. The data centers are not always in sync.

Overall I thought it was a pretty good article.