Google search primed for 'Caffeine' injection
Google has completed testing on "Caffeine," a semi-mysterious overhaul of its back-end search infrastructure, and it will soon roll the new platform behind its live search engine. In mid-August, Google unveiled a online sandbox where it invited world+dog to test the new infrastructure, but as noticed by Mashable.com, the sandbox …
I ran a search on google just now and I got 9,380,000 results in 0.19 seconds. Why on earth would they want to spend money to bring that time down? The page took longer than 0.19 seconds to load and at this time in the morning my brain took a lot longer to process what I was seeing.
Although, I am so excited the day will be coming I will finally be able to run a sub 0.18 second search, my life will be so much easier and more efficient
"......the so-called GFS2."
That, from one who should know, would indicate to me that it's actually got another name internally, so we don't even know WTF the thing's really called, let alone how it really works.
Google are beginning to make the Stasi look open and approachable.
"So, somewhere in the world, there's a mega data center on the verge of sabbatical. Or perhaps it's already happened. Presumably, Google will tell us at some point. And tell us very little." I would agree, Cade, with the probability of it already having happened, and now up and running with Novel Protocols and Parameters ..... for Replacement of Search Information with the Placement of Discovery Evidence, thus to Morph and Transform Global Inquisition into Specific Intelligence.
And you just can't fault Google whenever it shares its Crown Jewel Gems Methodology so succinctly ..... "If you don’t know what’s going on, you can’t do decent back-of-the-envelope calculations!" ...... Page 29/73 in the recent Jeff Dean Google Fellow presentation (PDF) at a distributed-computing shindig in Montana....."Designs, Lessons and Advice from Building Large Distributed Systems"
* A much Softer Image to Present to Second and Third Party Parties rather than causing any undue alarm and unwelcome attention with a Hard Core In your Face Phreakery descript.
"a quintillion bytes of storage"
Or, to put it in terms that might mean something to somebody, that's a million terabytes, or 100,000 Libraries of Congress (is that an official Reg measurement? I couldn't find the guide to check...)
Also, @ Dave 160 - I don't think Google care much that you get your search results one hundredth of a second quicker, but I think they'll be pretty pleased if they can serve both your search results and the huge amount of streaming data from sites like YouTube in a more efficient manner.
What Google need to do is deliver the search results that I find useful, instead of pushing a load of advertised links.
I've lost complete faith in Google to find what I'm looking for and I've switched my primary search engine to another source ... not Bing, beofre you ask!
Why should some new site that has payed money, be ranked higher than a site that has a decade on the air and more reputation? What is the point of a free and level-paying-field Internet if it is just subverted somewhere else along the way?
I don't mind swallowing adverts with free servies, but when I search for information I want the results presented to me, to be based on something that has a little more intellectual value than cold, hard cash which has been spent with a view to getting yet more, cold, hard cash.
I know - anyone would think that Google was interested in expanding in the future...
...when a spanner in the works is a good thing.
Yeah, damn those Google workers for not opening the doors on how their expensively developed systems work in minute detail! It's almost like they want to protect their trade secrets?!
Why does that set alarm bells ringing?
Caffeine today, methedrine tomorrow. All your data belong to Google - and sooner than you think.
Aren't they talking about how quickly documents appear in the indexes, just as much as how quickly search results are returned.
I love the fact that Google are prepared to keep trying new things and advancing, rather than waiting for others to catch up. What engineer wouldn't want to work in that sort of environment!
Google has been driving me nuts in recent months since its change to not include everything in your query by default. For (purely hypothetical) example, if I were to search for 'donkey dwarf porn', I might end up with results that relate only to dwarf donkeys, with no mention of porn anywhere. This means I'm forced to change my query to the much clumsier '+donkey +dwarf +porn', or in some cases '+"donkey" +"dwarf" +"porn"' just to make sure it doesn't go looking for any variations of any of the words (like 'ass', for example- which completely and disturbingly changes the meaning of my search).
I hate being second guessed- if I put search terms in, I put them in for a reason. I'd rather have no results at all than just the wrong results of 'dwarf porn', 'donkey porn', or 'dwarf donkeys'. Sometimes you just need a really specific result, and Google is starting to fail me.
That leaves the other three food groups to be covered later: fruit, alcohol and icecream.
Is it time for the problem of organising the worlds information to be broken up, shared out and distributed across a wide range of organisations and computing platforms / devices. In the interests of privacy, competition and even from the perspective of power consumption this starts to look beneficial - having a single corporation (or even a small group) mediate access to the web's index is not a good solution.
@ Michelle Knight - I know what you mean. Google's methodology to web search is anchored on a 1990's situation where a web populated by academics, intellectuals and technologists shared information openly and referenced one another via HTML hypertext links. This is largely not the case today (ironically google's success has contributed to this). They face many challenges to adapt now to the changing terrain - as more information disappears behind pay-walled gardens, into specialist archives, directories, libraries and resources.
My best guess is that the internet will ultimately do for google what the x86 PC and Windows did for the IBM mainframe - the current trend back towards the mainframe (albeit the distributed x86 hosted data centre centric mainframe) will once again swing back to a topology based around a grouping of interconnected peers.
Initial observations reported that the subject became an highly animated, hyperactive, gibberish spewing mess. Later mood swings were reported as the subject became shaky, paranoid, depressed and finally unable to sleep.
You mean, like that library hologram in The Time Machine?
"New... powerful... hooked into everything, trusted to run it all. They say it got smart, a new order of intelligence. Then it saw all people as a threat, not just the ones on the other side. Decided our fate in a microsecond: extermination."
Mine's the hobo overcoat concealing the sawed-off Ithaca 37...
> if I were to search for 'donkey dwarf porn'
NOTICE TO @Anonymous Coward (CITIZEN no#11020a65f7d8c7e)
You have just triggered a level 2 IMP deep packet inspection ALERT.
A platoon of riot cops are presently being dispatched to ransack your home. A reporter from newspeak corporation is preparing the story that will ruin your public reputation for immediate wire release in multiple versions re-hashed for tabloid, broadsheet & TV audiences. Your partner and friends web experience will now bombard them with subtle hints that you are a pervert. Your employment and financial records have been flagged to prevent you from ever performing any meaningful role within this society ever again.
Regards, Jaqui & Alan.
"Why should some new site that has payed money, be ranked higher than a site that has a decade on the air and more reputation? What is the point of a free and level-paying-field Internet if it is just subverted somewhere else along the way?"
Why wouldn't you be going direct to said reputable site rather than asking Google to take you there?
Tried using it for a few weeks. Gave up. Old Man Google infinitely superior. Hope the release is being done very slowly and far, far away.
ARM+GPU = dream super?SC11 NVIDIA: 'Yup, we're on it'
Nvidia's Maximus: Hard-core 3D graphics on speedBlog Quadro-Tesla mashup for animation acceleration
Get your bi-weekly
HPC newsletter
HPC Analyst Blog
Whitepapers, video and audio resources
Reg Forums: Join the conversation
