VMware NSX: What we learned in 2014; what to look for in 2015
Before I launch into some predictions for 2015, I want to take a quick look back at the post I wrote a year ago, predicting how the network virtualization landscape would shape up in 2014. This is not just to pat myself on the back for getting a few things right (my predictions weren’t that bold in any case) but to see what the previous year can tell us about what’s coming next.
For starters, I predicted that we’d see more adoption of NSX both numerically and in terms of a broader range of customers. The numerical data has been impressive: at the end of Q2, Pat Gelsinger announced over 150 paying customers for VMware NSX, and by the end of Q3, that number had grown to more than 250. Furthermore, the adoption has definitely spread into a wide range of industries and customer types, with enterprises as diverse as Best Buy, USDA and WestJet joining the service providers, Web 2.0 companies and financial companies who were early adopters.
Speaking of service providers, I also predicted that we would hear a lot about NFV, which was true as far as it went. We heard a lot about NFV in the press with analysts and at conferences (I even talked about it at VMworld), but we still seem to be some time away from any large scale deployments. I don’t think we’re going to see a lot of NFV deployment at scale in 2015 either, but we will see some small-to-moderate scale trials or pilots.
And while I did suggest that higher layer (L4-L7) services would become more important in 2014, I have to admit that I underestimated just how important security services would become. Micro-segmentation, leveraging distributed firewalling and the lightweight creation of isolated virtual network segments, has become the driving force behind a significant percentage of network virtualization deployments. In fact, it’s fair to say that 2014 was the year in which security became the dominant business problem driving network virtualization. My colleague Brad Hedlund has written a great post on how powerful the NSX security capabilities are.
Speaking of business problems, one striking shift in the customer landscape in 2014 was that our discussions were much more focused on solving business problems rather than on the technology per se. Instead of customers saying “I want to hear about your network virtualization (or SDN) technology”, they were saying “I need a solution to the following business problem”. When I walk into a customer meeting and the first thing I hear is “we need to address security for east-west traffic in the data center” (which we do hear frequently), I have much more confidence that the conversation will lead to an NSX deployment than I do after a purely technology-focused discussion.
So what about 2015 then? I’d love to say that we’re going to find some new use case that will be as compelling as security has been in 2014, but I don’t really consider that a high probability event. In fact, I think the existing use cases (agility, self-service IT, micro-segmentation) will continue to drive adoption. Instead, I would categorize 2015 as a year in which we’ll see a focus on implementation, in three main areas.
First, you’ll see a very strong focus on execution by the NSBU engineering team to bring new advancements for the VMware NSX platform to the market. Second, we will see our customers increasingly moving to implement production deployments of NSX. And third, we’ll see more of our partners delivering GA (generally available) versions of products that are integrated with NSX as we enable and expand our partner ecosystem.
The increasing amount of production deployment of NSX will directly influence where our focus lies on the engineering side. That means focusing on somewhat “boring” issues like manageability, “day 2 operations”, non-disruptive upgrades, etc. It also means that, as we add new features and capabilities, we will of course continue to maintain the focus on high availability and scalability that has been a hallmark of NSX from the beginning. Fortunately we have a lot of experience with these aspects of network virtualization and we’re building on a solid architectural foundation.
So while this may not seem as exciting as the big architectural shift that we started with NSX in 2013, or the creation of a new “killer use case” that we saw in 2014, I still think there is a lot to be excited about for 2015. Implementation of network virtualization at scale, in production, with hundreds of customers and dozens of partners, is going to change the world of networking, just as we predicted years ago.