Google’s Data Center Strategy Questioned

August 10, 2009

Google fired up its engineering engines in the period between 1996 and 2002. As the company entered its run up to its initial public offering, Google had locked and loaded on some core principles. I am sure you have internalized these by now. It has been more than a 11 years since the Google came on the search world’s radar.

The Register’s headline “Will Google Regret the Mega Data Center?” raises an interesting question. The story was written in August 2009, more than a decade after the GOOG launched itself. Can decade old technology remain viable in today’s wild and crazy technical world? Cade Metz reported:

In the wake of Microsoft’s decision to remove its Windows Azure infrastructure from the state of Washington – where a change in local tax law has upped the price of building out the proverbial cloud – the company’s former director of data center services has warned that Microsoft and other cloud-happy giants may soon find that the mega data center isn’t all it’s cracked up to be. “[Large cloud providers] are burning through tremendous amounts of capital believing that these facilities will ultimately give them strategic advantage,” reads a blog post from Mike Manos, who recently left Microsoft for data center outfit Digital Realty Trust.

Yikes! Google. A fat and out-of-step dinosaur?

fat dinosaur fixed

Google has three dozen data centers, a model that Microsoft has emulated. Google has according to chatter about a million servers humming along. What is The Register’s take on this important issue? You will have to read Mr. Metz’s article.

My view:

  1. Google’s core technology is getting long in the tooth, but so far, that technology has proven itself to be scalable and extensible. Googlers think away each day looking for angles. So far, so good.
  2. Microsoft’s approach has been to emulate Google. The idea seems good but Google uses the sort of gizmos you can buy at Frye’s Electronics for the most part. (There are some big exceptions, however.) Microsoft, on the other hand, uses name brand hardware. There’s a cost issue that has not been fully understood even by the legions of azure chip and blue chip consultants romping in this field of inquiry.
  3. Newer solutions work where infrastructure is up to the task. The future, in my opinion, is going to look more like a blend of old and new.

Good question. The challenge will be answering it because the Google has an almost insurmountable lead in the Web search space. As I reported last week, Google has a new appetite for the enterprise. For that venture, the Google has some publicly disclosed but largely unreported technical innovations that will give it a cost advantage in that sector.

Googzilla does not show signs of arthritis yet. She is a fat and less agile Googzilla than she once was.

Stephen Arnold, August 9, 2009


Comments are closed.

  • Archives

  • Recent Posts

  • Meta