Seven Questions for Adam Selipsky, VP at Amazon Web Services
A little more than a week ago, I made a lightning-quick trip to Seattle to visit with a few companies based there. (More on that in the coming days.) One of them was Amazon.com. I stopped by the company’s still-like-new headquarters in the South Lake Union neighborhood to talk with Adam Selipsky, vice president of Product Management and Developer Relations at Amazon Web Services. I had lots of questions about cloud computing and the plans of the company that’s most widely associated with the phrase.
Given how much people talk about cloud computing and the number of companies that use it, you’d think the cloud was a huge business for Amazon Web Services–or AWS for short. It’s not. Amazon doesn’t break the units results out specifically, and instead lumps it into the “other” category for financial reporting purposes, which amounted to $953 million in 2010, or about three percent of Amazon’s total sales. But you can get sense of its growing importance elsewhere within its financial reports: Capital expenditures in 2010 were $979 million, more than 2.5 times the amount it spent in 2009. Amazon didn’t detail exactly how it spent that money–some went to offices, some to infrastructure for AWS–but it did say this: “We expect this trend to continue over time.”
Next week, Amazon Web Services will hit its five-year mark, so with that milestone in mind, I took the opportunity to ask Selipsky first about its beginnings, before diving into its future.
NewEnterprise: Adam, I remember vaguely the day that Amazon announced it was getting into the vaguely worded “Web Services” business, and really scratching my head at it. How did all this start?
Adam Selipsky: We built it out of necessity. We wondered why projects were taking longer than it seemed they should, and so we did a study and found that our engineering teams were spending about 70 percent of their time on non-value-add work like provisioning and managing their server and IT infrastructure, and not innovating on behalf of customers. That was a big aha moment. We got a blinding glimpse of the obvious, we realized we were not the only ones with this problem. We figured out there was a broad need for these basic technology infrastructure services.
When I think of all the companies that I know have started out using AWS, my head spins. Who’s your biggest customer?
We wouldn’t break that out. Zynga is one the biggest, obviously. Look at their growth. They couldn’t have grown the way they did without AWS. Netflix is another highly significant one. They said they have moved most of their infrastructure to AWS. They have a talented set of engineers who could build and manage their own infrastructure if they wanted to, but the point is that they don’t want to. They’d rather work on Netflix rather than the underlying infrastructure.
You’ve recently added a service called Elastic Beanstalk. Are you going to continue to add services like that through the year?
As fast as we knock out new services and features, the list of what our customers want from us continues to grow. I think that’s because this is so new, and since we’re replacing the data center, there’s a lot of things built up over the decades that run in those data centers. There’s going to be a long list of mission critical services that we need to bring up.
What’s the big one that people are asking for?
More flavors of databases. We have RDS, the relational database, and we’ve said we’re going to do an Oracle engine. So there’s more to do on databases, not only adding new flavors, but also just being more scalable. There’s big demand for really highly scalable, high-performance databases, and I think we’ll continue to work on that. And I think that there will be more stuff around ease of use. There’s going to be more features to help you figure out and deploy whatever you need.
We’re starting to see a lot of talk from other companies like Google and Microsoft and IBM who want to get into the cloud services business. What kind of competitive threat to see coming at you?
We’ve been in business for five years with real paying customers, with large companies down to guys in basements, and I think we were lucky enough to have a real first-mover advantage. A lot of technology companies have not followed quickly in part because they haven’t wanted to follow quickly. They make money by getting seven-figure purchase orders, which is good for them, but not so good for their customers, so they have a real dilemma about making the transition.
Another thing I’m hearing a lot of these days is talk of private clouds. Companies are wondering why they can’t build their own clouds themselves, and still maintain the on-premise control of their data. Do you sense any threat from that?
There are a lot of companies with large sales forces who are banging the private cloud drum very loudly. When they all do that in concert, some customers are going to listen. But you have to parse that talk from the reality. I think some customers have to be very careful because they lack some fundamental characteristics of the cloud. If you’re still writing a very large check, and not operating in a pay-as-you-go fashion, you’re losing one of the key advantages of the cloud. If you just bought machines that had to appear on a loading dock somewhere, then what you have isn’t instantly scalable. Most importantly, it doesn’t take the heavy lifting out of your hands so you can focus on the things that your customers care about and instead focus on things that are really just table ante at the end of the day. CFOs and CIO are going to have look really closely at the benefits that are being advertised to them. A lot of those models look like yesterday’s economic model.
What about hybrid models? Along with private clouds I hear a lot about mixing cloud and on-premise.
I don’t want to come across as some kind of religious zealot for the cloud. We’ll live in some kind of hybrid world for many years to come. But I think that what hybrid principally means is when there’s cases where there’s applications that can’t move because of government regulations or it was just built in 1978 and it’s just not conducive to moving–you see that a lot actually–but you want to run other things in the cloud. One of our jobs is to make it really easy for enterprises to live in that hybrid environment.