Understanding Amazon Web Services

There are two ways to look at Amazon.com: as a retailer, and as a software company that runs a retailing application. Both are accurate, and in combination they explain why Amazon, rather than a traditional computer company, has become the most successful early mover in supplying computing as a utility service. For Amazon, running a cloud computing service is core to its business in a way that it isn’t for, say, IBM, Sun, or HP.

In a brief but illuminating video interview with Om Malik, Amazon CEO Jeff Bezos underscores this point in describing the origins of Amazon Web Services. “Four years ago is when it started,” he says, “and we had enough complexity inside Amazon that we were finding we were spending too much time on fine-grained coordination between our network engineering groups and our applications programming groups. Basically what we decided to do is build a [set of APIs] between those two layers so that you could just do coarse-grained coordination between those two groups. Amazon is, you know, just a web-scale application.” As it developed the APIs for its own applications developers, it realized that the interfaces would be useful, as well, to other programmers of web apps: “And so, look, let’s make it a new business. It has the potential one day to be a meaningful business for the company, and we need to do it for ourselves anyway.”

Bezos goes on to note that Amazon’s retailing operation is “a low gross margin business” compared to software and technology businesses, which “tend to have very high margins.” The relatively low profitability of the retailing business gave Amazon the incentive to create a highly efficient, highly automated computing system, which in turn could become the foundation for a set of cloud computing services that could be sold at low enough prices to attract a large clientele. It also made a low-margin utility business attractive to the firm in a way that it isn’t for a lot of large tech companies who are averse to making big capital investments in new, low-margin businesses.

“On the surface, superficially, [cloud computing] appears to be very different [from our retailing business],” Bezos sums up. “But the fact is we’ve been running a web-scale application for a long time, and we needed to build this set of infrastructure web services just to be able to manage our own internal house.”

This is the great advantage that, at this early stage in the evolution of the utility computing industry, is held by companies like Amazon and Google. Building an efficient cloud-computing infrastructure does not represent an added expense for them; it’s a prerequisite to the success of their existing business.

5 Comments

Filed under Uncategorized

5 Responses to Understanding Amazon Web Services

  1. Ed Anuff

    Amazon is taking a long term bet on how applications will ultimately be bought and paid for in the cloud-computing future. The idea of reselling their computing capacity as a utility was a great idea, but where it starts to get very exciting is when you start to imagine the post-Web 2.0 world where rather than sharing resources with thousands of other users with no consistent quality of service, you’re instead able to purchase with one-click, and have immediate access to, your own private installation of a web application with a guaranteed level of computing and network resources for your company and your customers. While this may seem retro to many, the current multi-tenant model for software-as-a-service assumed that most people were not making enough usage of their applications for dedicated resources to be necessary. Multi-tenancy will still be the norm for free consumer services, but as web applications increasingly become accepted and preferred for all sizes of business, more options that provide reliability and consistent quality of service will become required and demanded. I believe that this is quickly going to lead to a “shrink-wrap 2.0″ movement emerging in the software industry which is going to greatly expand the number of applications available as services, and Amazon is going to be the retailer of these applications, making them, in the long run, the greatest competitor to companies like Salesforce.com.

  2. >> There are two ways to look at Amazon.com:

    >> as a retailer, and as a software company….

    A “hard core” systems analysis would examine the physical flow of widgets and funds through the economy to determine what kind of business model best describes Amazon. Amazon doesn’t look much different than the catalog mail order businesses of the late 19th century. I’m not saying Bezos and friends aren’t clever and innovative in putting a web interface front end to it, but credit needs to go to the guys like Franklin, Ward and Sears who actually invented the cataloged/remote purchase business model. BTW, wasn’t Sears involved with the first ISP Prodigy?

  3. I’m one of the bigger AWS fans out there, but one thing that concerns me is their long-term prospects against native platform companies that have a *lot* to lose if Amazon becomes the dominant player in this market. I did a quick post about this here: http://tinyurl.com/5dnwrw

  4. I’m just starting to get into cloud computing and especially could databases. The company I work for, enterprisedb, is also moving into this space. The one issue I still see is SLAs. The current level of guarantee is ok for many companies but most larger companies don’t want a credit next month, they want the system up.

    Of course, if they’re large enough, they will probably never even consider moving mission critical apps outside their own data center.

    Even with that though, for development and testing, the cloud is awesome. Each development team and testing team can have it’s own virtual area. Need an additional dev box? Don’t buy one, just add one online.

    Thanks,

    LewisC

  5. LW

    It’s interesting to see the comparison between Amazon’s Elastic Cloud Computing service, EC2 and Salesforce.com’s force.com platform. There are huge advantages of both platforms as they address very different approaches to solving business problems. One advantage of deploying applications on the Salesforce.com platform, Force.com, is the built in functionality, for example, reporting, light analytics, and work flow rules. In essence, you have the ability to build a very sophisticated application leveraging the force.com in a matter of hours or days. Amazon’s platform could certainly be used to build a stellar application, even a multi-tenant application and it’s uniquely different than Salesforce.com because it’s a core infrastructure solution to provide computing power on-demand. for the ISV or IT department who has already built a server based application, they can leverage EC2 to expand computing power on demand as peak usage climbs or shrink computing power as demand declines on non-peak usage periods.

    IT Departments have a significant advantage moving forward with both an application platform, salesforce.com and an infrastructure platform, that can assist in quickly solving business problems and core infrastructure elasticity.