I attended last December’s North America KubeCon where the Network Service Mesh project had its first big public outing, showing a proof of concept demos. Network Service Mesh addresses some of the network level inefficiencies that arise with modern virtualization infrastructures. In cloud native environments, networking runs by default at a relatively high level of abstraction.
But some workloads—telco workloads, for example—run much more efficiently when they have access to lower networking levels. Network Service Mesh abstracts the network links so that any particular workload can run at the networking level it’s best suited for.
I’ve spent over a decade in networking of one kind or another—from embedded devices to big network servers—and have long been interested in Software Defined Networking, Network Function Virtualization and, recently, networking in the cloud native world. So, when Network Service Mesh was launched earlier this year, I was immediately intrigued and kept an eye on its progress.
Being part of VMware’s Open Source Technology Center, I was involved in other Service Mesh projects like Istio. However, last fall, I started looking closer at Network Service Mesh and joined the project with a number of patches. My first major contributions were helping prepare Network Service Mesh for its KubeCon debut. In close collaboration with the established project members, we got the demo up and running and did some stabilization work to have the project ready for its moment in the spotlight.
We ran two Birds of a Feather sessions at KubeCon: an introduction and a deep dive. At the end of the intro session, I was thrilled to be named as an official contributor to the project. It was great, too, to see the project receive wider attention. It had been known about in the open source networking community, but the KubeCon talks attracted around 200 people and seemed to generate interest in the broader open source community. Project contributors also had multiple positive hallway conversations afterward and we even had several potential large-scale users come up to us wanting to learn more.
Before KubeCon, we were mostly focused on proving the validity of the idea that became Network Service Mesh. But now that we’ve shown it can work, we have important testing and stabilization work to do, so that’s where I plan to focus my own contributions for a while. Looking further ahead, I’m also interested in thinking about how Network Service Mesh gets integrated into a wider ecosystem and helping make that happen.
Network Service Mesh was started by respected developers from Cisco and Red Hat (soon to be part of IBM) and already has more than a dozen other contributors. Now that VMware is backing it, too, I think that’s a strong message to the community that this is a project with real potential. If this all sounds intriguing, I encourage you to go to our home page to learn more about what we are doing and even begin making contributions of your own if you like. And be sure to stay tuned to the VMware Open Source Blog and follow us on Twitter (@vmwopensource) for more on open source Service Mesh solutions.