[LINK] Caffeine

[stephen at melbpc.org.au]

The Holidays are Over. Is it Caffeine Time? 

 http://archive.webpronews.com/2010a/0126.html


Last summer Google announced a new project called "Caffeine", which was 
described as a re-write of Google's web search architecture. 

Around that time, Matt Cutts discussed Caffeine, comparing it to the "Big 
Daddy Update" of 2005, which consisted of changes to the way Google 
crawls and indexes websites. 

It appears that more people are now seeing the effects from Caffeine out 
in the wild.

Back before the holidays, Google made it a point to assure everybody that 
Caffeine would not be rolled out (except for at one data center) until 
after the holidays were over - January at the earliest. The reason for 
this was that Google didn't want to shake everything up during a key time 
for businesses (they didn't want a repeat of the Florida update). 

The company let everyone know about its intentions at PubCon in November. 

In fact, a few days ago, Google's Matt Cutts posted a video running 
through his presentation from that event on his blog:

 http://www.mattcutts.com/blog/state-of-the-index-2009/

He also provided the slideshow. It covers much more than just Caffeine, 
but if you missed it, you may want to consider watching it anyway 
(Caffeine discussion starts at about 22:10 in the video and at slide 29 
in the presentation).

"It's a re-write of our indexing infrastructure. It's taking the old way 
that we used to index things that we'd crawled around the web, and we're 
replacing that with new architecture that's fresh and that had been 
written to be more scalable, more flexible, [with] the ability to attach 
different types of data, and in the process of indexing, the ability to 
do more documents for a more comprehensive version of the web, and the 
ability to do it faster," Cutts says of Caffeine. 

But enough background. Barry Schwartz at Search Engine Roundtable points 
to a WebmasterWorld forum thread www.webmasterworld.com/google/4053085.htm
where administrator Tedster claimed to have seen Caffeine in action at a 
number of IP addresses. He wrote: 

"I'm seeing the Caffeine data-set being served via this set of IP 
addresses: 64.233.169.147, 64.233.169.105, 64.233.169.103, 
64.233.169.104, 64.233.169.99 and 64.233.169.106"

It seems to take 5 IP addresses to build the complete SERP, where in the 
past it often took only 3. 

Schwartz also pointed to another member's post (Whitenight), who said: 

"Well, just tripled checked with offices/employees in Texas, Colorado, 
and Indiana. All 5 "control" keywords/sites showed live Caffeine."

That member's latest post www.webmasterworld.com/google/4053085-11-30.htm
says that the Caffeine Dataset is also on http://66.102.7.99 and 
http://66.102.7.104.

We don't know for sure if this is all really Caffeine in action though. 

Google hasn't commented on it, and has not made any announcements 
regarding Caffeine since what Matt said above. Some people don't believe 
this is Caffeine at all. As Schwartz notes, we'll have to wait for Google 
to say something. 

Still, January is almost over, and Google said it would wait until after 
the holidays, specifically mentioning the month of January. It's about 
time for this to be rolling out to some extent. 

Speed has been emphasized a significant amount in Caffeine discussion, 
and Cutts told us that page speed would likely become a ranking factor. 

Regardless of whether or not you are witnessing Caffeine in action yet, 
rest assured that it will be here sooner or later, and any edge you can 
give yourself in the meantime is for the good of your own site's 
performance. 

Speed will not only supposedly help you in search going forward, but it 
just makes for a better user experience. 

--

Cheers,
Stephen