Last week Snowflake was at the Amazon Web Services re:Invent conference in Las Vegas, which has quickly become a huge show of all things in the cloud. For the first time since our October announcements, we publicly talked about and demonstrated what we’ve been working on for over two years. We had a huge amount of traffic stopping by our booth, not just to check out our giveaways (although our scratchers and prizes were a huge hit) but also because people were interested in understanding what technology could help them process data in the cloud. We had a lot of great conversations with people who were actively looking for solutions because they are already seeing more and more of their data coming from or being stored in the cloud.
One theme that struck me from hearing the speakers and attendees was that cloud has become more than just infrastructure for rent—it has emerged as a new platform for building software. Having been a developer myself in the days when client-server was the dominant paradigm, I found it particularly fascinating to see how the approaches that people are taking in the cloud today are radically different than the way that we developed software in earlier paradigms. Hearing users at the conference talk about how they were building various applications in the cloud, I thought of how much faster and easier it had become for users to do that.
The same paradigm shift was on display when it comes to data processing. Just deploying a single component of the data pipeline used to be a huge undertaking—I remember days spent just to set up an application server, let alone to get the software running on top of it integrated with all the other components in the pipeline. Now people talk about constructing data pipelines on the fly with a small amount of glue needed to connect various components delivered as services—what used to be a multi-month project just became a live exercise in a one-hour conference session.
That’s a brand new world. I’m looking forward to what’s going to happen next.