Broadcom is bringing to market a pair of Kubernetes-focused upstream open-source efforts targeted at providing a more consistent platform for developers and also extending the vendor’s long-standing work with the Kubernetes and open-source community through its own operations and those of its VMware operations.

The first effort is the acceptance of its VM API project into the Cloud Native Computing Foundation (CNCF) “Sandbox” environment, which is the entry point for open-source project entering the CNCF ecosystem. Broadcom positions the project as a way to allow virtual machines (VMs) to be provisioned and managed through the Kubernetes API across multiple heterogeneous infrastructure providers.

Prashanth Shenoy, VP of product marketing for Broadcom’s VMware Cloud Foundation (VCF) division, explained during a press briefing that this work is based on the vendor’s view that “virtual machines should be represented as [a] first-class Kubernetes resource.”

Shenoy noted this work is about preserving embedded VM characteristics as they are moved across different infrastructure environments, similar to what the Kubernetes community has done with its Cluster API project.

“How do we provide a common, vendor-neutral and provider-neutral API that allows organizations to create and manage VMs across different environments while retaining the capabilities of the underlying infrastructure,” Shenoy said. “This becomes pretty relevant for demanding workloads, including AI, where the performance characteristics, I/O characteristics, can be very crucial.”

Workload conformance builds on history

Broadcom is also co-chairing the newly formed Workloads Conformance Working Group within the Kubernetes community, a move that Shenoy said was focused on “providing that consistent expectation of how workloads should behave and run across diverse Kubernetes environments.”

Shenoy explained that the goal of the group was to ensure that already established conformance programs can product a level of production consistency between different operating environments.

“We feel it's an important opportunity for helping us to reduce this fragmentation and giving organizations greater confidence in workload portability,” Shenoy said of the group. “If they want to move their Kubernetes workload from a private [cloud] to a public [cloud], or a public [cloud] to an edge, from one distribution environment to another, they can [rest assure] that the workload will perform the way it's supposed to perform.”

Broadcom tied its interest to work it has done with the Sonobuoy project that is the diagnostic platform behind the CNCF Certified Kubernetes Conformance Program. VMware inherited that Sonobuoy work when its purchased Heptio in late 2018, a vendor that was founded by Joe Beda and Craig McLuckie who were involved at Google on its Compute Engine and the platform that eventually became the open-source Kubernetes project.

Shenoy framed the pair of updates under Broadcom’s view that, “at a high level … we strongly believe at Broadcom that Kubernetes, wherever they run and who ever distributes that, should be very consistent,” a belief that “requires a very consistent expectation for the workloads, common APIs, and frankly, continued collaboration across the open-source community.”

The work also continues Broadcom broader work in the cloud-native community.

The vendor earlier this year became a CNCF “Platinum” member, which is achieve by an increased investment into the cloud native-focused organization. Others in that tier include Amazon Web Services (AWS), Apple, Google Cloud, Microsoft, Nvidia, and Oracle.

Broadcom at that time tied the “expanded investment in CNCF” to “its ongoing commitment to open-source collaboration and the technologies enabling organizations to build and operate modern application platforms,” with a specific mention of enterprises increasingly moving “AI and data-intensive workloads into production.”

Broadcom has also been a long-term contributor to the CNCF ecosystem, including as one of the top contributors to the Kubernetes project, and has originated other projects like Velero, Harbor, Antrea, and Contour.

VKS support expansion

The CNCF late last year also bestowed a Certified Kubernetes AI Conformance Program gold star on Broadcom’s vSphere Kubernetes Service (VKS). That program was established to level-set expectations for AI workloads running in Kubernetes-managed container environments.

Broadcom this week noted that it has recently validated 20 new partners to its VKS program. This includes observability products from Cisco Splunk, Datadog, Dynatrace, New Relic, and Grafana Labs; security and compliance features from Crowdstrike Falcon, HashiCorp Vault, Sysdig Secure, Tigera Lynx, and Wiz; and data-intensive and AI workload solutions like Canonical Apache Spark, H2O AI Cloud, and SAS Viya.

“The idea is to make sure we work with these partners, help them validate the solution either jointly or provide a test bed and engineering resource on which they can validate their software or their service or their tools to make sure it runs on VKS,” Shenoy said. “This gives, ultimately, confidence to our joint customers and provides very clear deployment guidance for our customers that they can confidently run those services on top of VKS.”VKS support expansion