We held Network Service Mesh’s (NSM) first-ever NSMCon the day before the most recent KubeCon in San Diego. From my perspective as an NSM maintainer, it was everything we had hoped it would be and more.
If you aren’t familiar with the project, NSM is an open-source hybrid/multi-cloud IP service mesh that provides “missing” Kubernetes networking capabilities via a simple set of APIs. It is a sandbox project within the Cloud Native Computing Foundation.
NSMCon offered a full day of presentations, a community lunch, and then a speakers’ panel to wrap things up. You can find video recordings of all of the sessions at the NSMCon event page as well as most of the slides people shared.
It really felt like a watershed moment for our community. As I said in my maintainer’s keynote, a year ago I would never have guessed that we’d be ready to host a conference on this scale. But we ended up around 80% oversubscribed and were forced to turn people away.
Quite a few newly-launched projects were presented in public for the first time that day, and we had people from a wide range of companies both speaking and participating actively in the discussions. I left with multiple new contacts and several exciting new projects to discuss and plan for the future.
My own talk aimed to articulate “why NSM matters.” Essentially, I described the kind of architects and engineers that we’re hoping will engage with NSM, the tools that we’re looking to provide them, and then some specific examples of what NSM might enable them to do. If you want to know more, check out the talk here. Later in the day, I shared the stage with Taylor Carpenter, project lead for the cncf.ci + CNF Testbed, to talk about facilitating more complex use cases in the CNF Testbed with NSM. And I want to give a shout out to my VMware colleague Ivana Atanasova, who spoke about enabling traffic observability and metrics within NSM. You can find her talk here.
One of the most gratifying things at NSMCon was seeing all the “behind-the-scenes” action, where people were talking with each other about what NSM can do for them and how it might integrate into their own products and projects. A lot of these conversations and expressions of interest featured people we hadn’t heard from before in our weekly group calls, so it really felt like we’d reached a new level of interest in the project.
Some of these conversations went in surprising, and very positive, directions. There turned out to be a huge interest in a “virtual inter-cluster private L3 network” use case, for example, which was shared in one of the talks. Lots of people came up to us afterward and asked about it. We continued those discussions in the follow-up weekly group calls, and it got so big that we now are considering some API changes to better accommodate this use-case and similar ones that will emerge. My talk also connected me with Milind Gunjan from Sprint, who shared the company’s efforts with the open-source 4G Evolved Packer Core implementation in OMEC. We are now cooperating to bring this up with NSM, have submitted a talk for KubeCon EU, and are planning to get this implemented as a high-performance 4G example in the CNF Testbed.
The growing sense of interest in NSM was solidified throughout the rest of the week. NSM was the subject of both a KubeCon keynote and another KubeCon talk. During the larger conference, we also met with people from other CNCF projects as well as attendees who wanted to follow up with us on ideas and projects that they’d learned about at the NSMCon Day Zero event. It was great to find such a wide range of people who were aware of what we were doing and curious to learn more.
That rich set of connections and interactions really speaks to the value of running a project-specific, in-person event – so we’re hoping to do another NSMCon at KubeCon Amsterdam in the spring, coupled with an even greater NSM presence at KubeCon itself. Hope to see you there!