Another Times piece on the Google phenomenon, with some interesting facts. Paul rounds them up.
The focus of the piece is on how Google has made infrastructure a core competence, to the point of designing and building servers and related software. This is not news to readers of Searchblog, but there are interesting contours to the story, including a response from Yahoo:
“At some point you have to ask yourself what is your core business,” said Kevin Timmons, Yahoo’s vice president for operations. “Are you going to design your own router, or are you going to build the world’s most popular Web site? It is very difficult to do both.”
Google, in fact, has decided it will do both.
And from Microsoft:
“Google doesn’t have anything magic here,” Bill Gates, the Microsoft chairman, said in an interview. “We spend a little bit more per machine. But to do the same tasks, we have less machines.”
Google clearly believes that its infrastructure will prove a lasting competitive advantage, and the piece hints that Google may even get into microchip design. Hmmm, is Intel paranoid now?
7 thoughts on “NYT on Google Infrastructure”
“We don’t think our competitors can deploy systems cheaper, faster or at scale,”
I’ve never used a Yahoo service that has failed on me like Google Reader, Google Analytics, Gmail, and other Google properties that have buckled under the pressure of high traffic.
I did hear of an engineer working at Google whose previous job was designing some of the on-chip logic alongside an embedded ARM core…I think the Google Switzerland office should be closely watched!
The Yahoo point is an interesting one, but in my opinion Google succeeded in doing both things. Services has failed sometimes, that’s true, but now they learnt their lesson; that’s what “invitation beta” is all about.
The Microsoft point is really out of the world. Yes they have less computers but Google always thought about performance/watt not really performance/space.
Entering the processor business for their own servers seems reasonable, if processor makers don’t produce the low power usage components they need they’ll have to start making their own. The ARM architecture has been thought for low power consumption in small devices; using it in servers is a good idea when you want to save power.
Readers who work for telcos will recognize Google’s dilemma, do you own the network or rent it? Hardcore Bellheads will insist on owning it to control it. Message to debt-laden telcos, when you have loads of cash and brains – it’s not a dilemma…
I think Google can supply free services and occasionally buckle under the load and people will be forgiving. Sometimes my Gmail is delayed – does not happen often but things like that do happen – but they give us so much – I feel overwealmed with the services Google is providing us.
What’s the limit of the infrastructure? Probably none – the only thing I’d look at – if I were Google – is it takes a different mindset to engineer something vs. running it well. Sure – if you design it right it ought to run itself…and that’s what they aim for – but some things can’t be totally automated – and businesses also need human beings to run them – not all of this can be engineered.
The challange that Google has – is not that it can’t do both things (as the article suggests might be a problem) but how well it can do both things. Right now – Google goes after innovation at the cost of running the businesses/products they create – which they don’t have the focus to run and increase marketshare.
Google getting into CHIP design is totally baseless. A Redmond,WA based hardware manufacturer located just off 520, has supplied hardware to Google worth $1.2 billion in the last 8 months.
It’s a bit presumptious of Bill to assume they do the same tasks as Google, evidently given the fact that his search engine is a low third in the running and swamped by spam they do not.