Trending: Regulation is coming: How startups should be thinking about data, security and privacy

Amazon likes to say that it focuses on customers, not competitors, which means it’s rare for the company’s top executives to talk about their rivals publicly. So a talk by Amazon Web Services CEO Andy Jassy this week at the University of Washington was notable in part for Jassy’s comments about the competitive landscape in the cloud.

Cloud Report Card: Amazon Web Services is a $12B juggernaut, but Microsoft and Google are gaining

University of Washington computer science professor Ed Lazowska, who led the discussion, got Jassy going on the topic by asking him for his theories about why Google and Microsoft have taken a different approach than Amazon in the cloud. Jassy went on to address topics including the prospects for Oracle and IBM in the cloud, the battle for cloud enterprise customers, and concerns among customers about vendor lock-in.

Continue reading for excerpts from Jassy’s comments on AWS and the competition, and watch the video above for his full talk at the UW’s Computer Science & Engineering department.

On the two different “categories” of cloud competitors

“There’s one category of large technology providers, like some of the ones you mentioned, who decided to build an abstraction much further up the stack. I don’t know exactly why they pursued that. I think some of it might have been just a feeling like the basic building blocks were too basic, that what people really wanted was something much more abstracted. We were very focused in the early days on developers and on startups, and we still remain incredibly focused on them, but we had this perception that they would be the ones who would try the platform first, their alternatives were much worse than enterprises, they tend to start and adopt any new paradigm shift like this was, and that group of people is willing to dig into the details and stitch together the components maybe in a way that enterprises weren’t as likely to do. So we started there, and we figured that enterprises would choose to have a bunch of people who did that stitching, and then for others who would want abstractions further up, we’d build the abstractions later when we had that adoption. I think some of the other folks, either because their customer base focus was more enterprise, and/or maybe because they thought the building blocks were too primitive and not interesting enough, started in another spot.

“I think there’s a second category of large technology companies that just took an ungodly long time to get there. These are companies like Oracle and IBM, some of those folks. I think for them the model that we were pursuing, in the cloud, was so disruptive to their core businesses. The margin structures are radically different. The pricing structures are dramatically different. The delivery model is radically different. The way you take care of customers is radically different. So different that I think you have that dilemma at a lot of large companies: do you really want to try and accelerate the cannibalization of an existing … I think that they fought as long and hard as they could. They pooh-poohed it and they said, first no one will use it, then maybe only startups will use it, and they won’t use it for anything real, then enterprises will never use it, then enterprises will never use it for anything mission-critical. Companies and developers voted with their workloads, and so now they’re in this spot of trying to spin something up. It’s six, seven years late.”

Amazon Web Services CEO Andy Jassy, right, speaks with UW computer science professor Ed Lazowska this week in Seattle. (UW via YouTube.)

What has surprised AWS as the cloud has evolved

“There’s so many things in our business that have been surprises to us. Probably the biggest of which is that it just took so long for the other large technology companies to offer something in this space, but I don’t think in our wildest dreams we ever thought we’d have a six- to seven-year head start, especially because we’re across the lake from another big technology company in that space, and we all know so many people at each other’s companies. And so, there have been so many surprises, but if you read the document, the stated mental model was that an individual in his or her dorm room or garage would have access to the same cost structure and scalability of infrastructure as the largest companies of the world. To give you an idea back then, we were talking about those largest companies being like, Peoplesoft. We wanted to be part of enabling a thousand flowers to bloom.”

How the competition will shake out

“If you look at the market segments that AWS is addressing, it’s infrastructure software, hardware, data center services, and then we’re building software further up the stack, that’s trillions of dollars globally. There’s not going to be one successful player. I don’t think there’s going to be 30, because scale really matters here, both with regard to cost structure as well as the breadth of services. But there are going to be multiple successful players, and who those are is still going to be written. But I would expect that several of the older-guard players will have businesses here, as they have large installed enterprise customer bases, and a large salesforce, and things of that sort. There’s a lot of work those companies have to do to find the type of functionality that you have in AWS today. We have a lot more functionality by a large amount than, whether it’s Microsoft or really any other provider. For reference, we also have a much larger set of security capabilities and compliance capabilities than you’ll find anywhere else.”

On the competition for enterprise customers

Amazon Web Services CEO Andy Jassy at UW this week. (UW via YouTube)

“What’s happening is most enterprises are making decisions around, How many companies do I want to use in the cloud? Do I want to move to the cloud? How do I want to do it? How many companies do I want to use? Most are predominantly deciding that they’re going to use a single provider. Many will have a second provider they use for a relatively small percentage of their workloads, so that they know they can move, should they decide they want to move, but the benefits of being with a single provider — around not having to standardize on the lowest common denominator, not having to have your development teams have to be expert on multiple platforms, and then leveraging your buying power across one provider instead of splitting it across two or three — are just too great when you actually get into the practicality than to split it across a few. So most are predominantly picking an infrastructure provider, and a lot of what matters to them at the end of the day is the breadth of capability that’s there, because it not only allows them to build any new application or idea they have, but most of them are looking to migrate their existing applications over the next few years, and having the right tool for the right job allows you to be much more cost-effective and agile, which is what a lot of them are trying to do.”

On the risks of getting “locked in” to one cloud provider

“Customers will always be nervous about lock-in, and I think the experience they had particularly with a company like Oracle, where it’s a really hard thing to get out of, and they’re so hostile to their customers, that I think it’s a concern for every enterprise. Almost always when I talk to enterprises about lock-in, that’s the example that they cite. They worry that they’re going to use a platform and suddenly start raising prices — even though we’ve lowered prices on 59 different occasions with no competitive pressure to do so in the last eight years — they’re going to always be sensitive about that, and they should be thoughtful about it. But the reality is it’s so much easier to move from the infrastructure, cloud computing platforms than it is from something like Oracle. The APIs look pretty similar. All the migration services we build allow people to move to us, or away from us, and back on prem, if that’s what they want to do. And the business model is different, where we’re not signing these five- to 10-year unlimited licensing agreements and then finding ways to audit people and keep them in the loop. They can leave whenever they want, and we know we have to earn the business every hour, every day, every week of the year.”

More from the UW CSE Colloquia series

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline


Job Listings on GeekWork

Find more jobs on GeekWork. Employers, post a job here.