In this exclusive VMblog Q&A, we sat down with Ian Knight, a prominent figure in the world of cloud-native technologies. During this discussion, we dive into the evolving trends surrounding the cloud-native journey, as well as the revolutionary concepts of eBPF and Cilium. As the CRO of Isovalent, a leading company at the forefront of cloud-native networking, Knight brings invaluable expertise and perspective to the table.
Join us as we uncover the key insights and practical knowledge shared by Ian Knight, shedding light on the transformative power of cloud-native architectures and the game-changing capabilities of eBPF and Cilium.
VMblog: You’ve been involved with containers and Docker for a large portion of your career–what are some of the trends you are seeing with enterprises today in their cloud-native journey?
Ian Knight: The biggest trend that I’m seeing in Kubernetes today is strong demand for multi-cloud capabilities. The concept of multi-cloud has been talked about for a long time, but it’s been mostly vendor hype, and mischaracterized in terms of simply connecting public cloud providers together.
What platform teams are really going for with their Kubernetes infrastructure isn’t just connecting public clouds–but more about achieving a unified architecture to connect existing on-prem infrastructure AND public and private clouds, while maintaining uniform security and observability layers. Platform teams want a single connectivity layer to unify all of their infrastructure.
VMblog: Why the general momentum for eBPF and Cilium right now? What are the needs in cloud-native modernization efforts that make these attractive technologies?
Knight: What’s evident today is that we’re “day 2” of cloud-native infrastructure. Platform teams have advanced requirements that don’t ship natively with Kubernetes and Docker–especially at the networking layer.
Networking never moved away from the device-centric model, when it was all about plugging hardware switches and ethernet, and there were network operators. The age of virtualization for SDN and virtual networks–all that did was make something done in hardware and implement it in software. But the network remained device- centric, and designed for the age of virtualization.
With Kubernetes, platform teams made a giant leap, where use cases go way beyond devices and servers, and the requirements are about connecting services and applications. This so-called new “connectivity layer” for cloud-native infrastructure is what platform teams are chasing today.
And the same way that Javascript made the web browser programmable, eBPF and Cilium are enabling technologies that make the operating system kernel programmable for cloud-native use cases. There is an entire wave of innovation with cloud-native networking, security and observability, made possible by eBPF and Cilium.
Cilium in particular removes the device-centric thinking of previous iterations of networking. It moves platform teams away from IP addresses as the way they think about services, certs and service identities. It changes the perspective from virtual networks, to a universal connectivity plane. And it allows platform teams to operate and tune a single network, rather than managing multiple networks. The two drivers for Cilium and eBPF are really (1) simplifying operating Kubernetes and containers, and then (2) the more advanced use cases in security and observability that simply aren’t possible until you operate your network as a single connectivity layer.
VMblog: What made you excited about Isovalent and the company’s mission, specifically?
Knight: I joined Isovalent because I believe we have the right team, technology and track record to deliver the next generation of cloud-native networking infrastructure for “day 2” Kubernetes. As the original creators of eBPF and Cilium, the two most popular open source projects that are disrupting cloud-native networking infrastructure–Isovalent is uniquely the pioneer in this space. The company has earned a large following from platform teams that are responsible for redefining networking, security and observability in a single solution.
For my whole career, I’ve been a huge believer in open source. With more than 15,000 GitHub stars, 12,000 community Slack members, CNCF project status, and massive adoption at global enterprises–Cilium is the OSS project that’s driving the evolution of cloud-native infrastructure modernization.
In addition to building Cilium, the most advanced eBPF-based solution in the world, Isovalent co-maintains the eBPF codebase upstream in the Linux kernel and is the leading company driving a rapidly growing open source community around eBPF. Isovalent is heavily involved in open source communities and has led the creation of the eBPF Foundation (along with founding members Meta, Netflix, Google, and Microsoft).
Enterprises that are modernizing their infrastructure for cloud-native want to work with the creators and maintainers of these canonical projects, because they are light years ahead of other networking players that are jumping into the space.
VMblog: What do you think the future looks like for cloud-native, at the network connectivity level? Where is this all headed?
Knight: Cloud-native demands a complex set of requirements from the infrastructure layer. Infrastructure software must be capable of running equally well in public and private cloud infrastructure. It must be vendor agnostic and open source. It must be highly scalable to meet the agility of containers and CI/CD. It must be highly secure because it often runs outside of company premises. And it must still meet the traditional enterprise networking requirements in terms of interoperability, observability, and security. In my time at NGINX, it was evident that open source software was disrupting the network stack and that Kubernetes networking and service meshes were on an evolution to merge into a combined layer. I believe that Isovalent’s leadership of eBPF will be the defining factor that enables the company to make this single connectivity layer a reality for enterprises.
VMblog: What happens with all of the existing virtual machine and other legacy infrastructure as enterprises move to cloud-native, from your point of view? Obviously there is a huge installed base of VM architecture–how do the new and old worlds meet moving forward?
Knight: The newly- launched Cilium Mesh capability in Cilium addresses this bridge between cloud-native and legacy environments. It brings a transit gateway to connect to existing networks. Any machine or network can plug in to Cilium Mesh via the transit gateway and connect to other machines or to Kubernetes environments natively. Cilium Mesh provides a single layer of connectivity that seamlessly delivers cluster-to-cluster Kubernetes connectivity (Cluster Mesh), fully distributed network segmentation, encryption, mutual authentication with mTLS, load-balancing at L3-L7 (Service Mesh and Gateway API) as well as mature Kubernetes Networking primitives (CNI) that have made Cilium the most widely adopted cloud native networking connectivity plane. In a matter of years, there will be no “cloud-native” versus “legacy” infrastructure–it will all be accessible through a single connectivity layer, based on Cilium and eBPF.
##
