Amazon’s Storage Rain Forest: Google Lost in a Jungle

August 21, 2008

Amazon is behaving more like a hungry group of jungle foragers than a company selling books online. In the last couple of years, Amazon has evolved with some zigs and zags into a store front with a services business. Most surprising to me is that Amazon doesn’t spend Google- or Microsoft-sized billions on its infrastructure if the company’s financials have been interpreted correctly by me. Last year, I recall figuring out that Amazon squeezed by on hundreds of millions for plumbing, research and development, and engineering. Amazon’s spending was interesting to me because Google and Microsoft reported spending billions on these essential activities. My original hunch was that Amazon had some magic blend of herbs and spices that gave its Amazonians super powers. I even explored clever techniques for getting more mileage from every technical potion that Jeff Bezos (arguably the smartest man in the world by some Amazonians’ accounts) and Dr. Werner Vogels (the second smartest man in the world by some of those same cheerleaders) assembled. I found tantalizing hints that graduate students from Europe were engineering on Amazon’s behalf, then sent back to the flatlands of the Netherlands to work outside the spotlights directed at Amazon. One person suggested that Microsoft and Amazon are “collaborating” but a question to me from a well placed Microsoft executive suggested that Microsoft is as curious about Amazon as some Wall Street mavens are.

The Amazon Lightning Stike

Once again, Amazon has sent electrical strikes through cloud computing. The company has introduced what amounts to a low cost Storage Area Network running on Amazon’s servers. An SAN makes it easy to set up a storage space and then forget it. In a sense, a user of this new service gets unlimited storage without the headaches of alternative methods.

A number of technical writers are enthusiastic about this new service. I want you to enjoy the fan raves yourself. Here are several of my favorite write ups that deliver PR manna for the Amazon tech gurus, Messrs. Bezos and Vogels:

  • Amazon’s own write up, stuffed full of juicy quotes about fault tolerance, engineering wizardry, and business pragmatics Amazon style.
  • Rightscale’s explanation, including useful diagram here. I suggest you snag the images and add them to your cloud storage clipbook. The pricing information is extremely useful too. The “availability zone” is a bit murky to me, but you may find it right for what ails your cloud needs.
  • GigaOM’s business analysis here. The key idea is that traditional data centers should be worried. Very worried. This well crafted article could have been crafted by a New York consultant carrying a Harvard MBA and a ThinkPad. As a former consultant, this is unusual in the world of “free” Web logs.

I have a different take on this announcement. Keep in mind that I am offering my opinion based on my research. If you want to read a slightly different take on Amazon’s jungle warfare prowess, stop reading. Proceed at your own risk.

My Take

In my forthcoming KMWorld column, I argue that cloud computing is useful in certain situations, but it is not yet appropriate for most enterprise applications. building a start up on an Amazon service is okay with me. A larger outfit can experiment with any Amazon service, and I will say, “No problem.” Ask me to put a mission-critical, regulated company’s applications and data on these services, and I will say, “Not me, sorry.” The problems are too frequent; the risks insufficiently documented for me. You may have a different view of walking barefoot through the foliage on the banks of the Amazon River. I’ve been there, and I am not going for a stroll no matter how innocuous and inviting the Amazon rain forest looks. My tutor in 1958 did this, and he ended up dead in a week. Reports of the death said, “No apparent cause.” That’s the Amazon rain forest in my view–risks you can’t see.

I have three other concerns as well. Let me mention them and ask you to read my KMWorld column when it appears in a few weeks:

  1. Stability. Anyone using MobileMe?
  2. Uptime. Anyone recall the Amazon Xbox ordering glitch or the mysterious all day outage this summer?
  3. Similarity. Anyone wonder why other vendors with arguably more infrastructure and more technical resources than Amazon have not offered this type of storage service?

For me, a company like Google should be in this business. Google is not. Is it because Google lacks the resources? Is it because Google doesn’t believe in the cloud computing revolution? Is it because Google doesn’t know how to offer this type of simple storage?

The answer to each of these questions is, “No.” Google itself makes use of these functions for its own services now, and it has for a number of years, based on my analysis of the company’s technical papers and patent documents.

Why?

This begs the question, “Why?”

My thought is that Amazon, like Microsoft, is in a hurry to out-Google Google. The path selected by Amazon seems to be to look at what Google is doing for its services and then offering a similar service to the developer world *before* Google takes action. Amazon stays “ahead” of Google. Amazon makes the GOOG look like a sluggish, overly conservative company facing increasing friction when it comes to innovation.

My hunch is that Microsoft is doing more than sitting on the sidelines and cheering Amazon on. Microsoft is learning and getting valuable assistance in the task of making Googzilla lame. I’m looking for more information on this cheerleading now.

Amazon is pointing the way to the way applications and other functions will work in a pervasive computing environment. My concern is that Amazon in its race to be first may find itself vulnerable when larger, better resourced organizations move into this sector as well.

The world’s smartest people at Amazon are confident they can win. I’m not so sure. I will avoid exploring the jungle without making sure of where I’m going and what I will confront. Agree? Disagree? Let me know.

Stephen Arnold, August 21 2008

Comments

2 Responses to “Amazon’s Storage Rain Forest: Google Lost in a Jungle”

  1. Speech-to-Text Analysis In Multimedia Content Discovery on August 22nd, 2008 2:49 am

    [...] Storage Rain Forest: Google Lost in a Jungle http://arnoldit.com/wordpress/2008/08/21/amazons-storage-rain-forest-google-lost-in-a-jungle/ Share and [...]

  2. Bookmarks about Infrastructure on September 28th, 2008 11:15 pm

    [...] – bookmarked by 2 members originally found by blume2000 on 2008-09-09 Amazon’s Storage Rain Forest: Google Lost in a Jungle http://arnoldit.com/wordpress/2008/08/21/amazons-storage-rain-forest-google-lost-in-a-jungle/ – [...]