Getty Images/iStockphoto

VMware Tanzu sets up edge computing showdown with OpenShift

Enterprise edge computing is still at a nascent stage, but a battle is brewing between Kubernetes platforms such as VMware Tanzu and OpenShift to support the emerging trend.

VMware Tanzu Kubernetes Grid 2.0 beefed up support for edge computing environments this week, catching up in some areas with rival Red Hat OpenShift and starting a fresh IT automation battle at the enterprise network edge.

Edge computing has yet to reach critical mass for enterprise IT organizations, but there are signs that interest is growing with 5G mobile networks, IoT applications and processing data for AI workloads where it resides rather than in a centralized location.

Gartner estimates that by 2025, more than 50% of enterprise data will be created and processed outside of data centers or public clouds. By 2027, machine learning will be included in more than 65% of edge computing applications, up from less than 10% in 2021, according to Gartner.

There's also some evidence that IT buyers are interested in connecting edge computing with existing cloud and data center infrastructure, where VMware Tanzu and OpenShift offer potentially relevant multi-cluster and hybrid cloud management features. In a May 2022 survey by Enterprise Strategy Group (ESG), 19% of the 387 respondents said application portability between data centers, edge locations and cloud services is critically important, and 67% said it was very important.

Meanwhile, it's still so early in enterprise edge computing adoption overall that there can be no clear vendor winner yet, said Gary Chen, an analyst at IDC. This gives VMware a chance to leapfrog competitors such as OpenShift as the market develops.

"I don't think anyone really has a head start in edge -- the market is very open and we see a lot of different vendors with different value propositions," Chen said. "VMware and Red Hat are pushing edge as an extension of the data center, while cloud providers are pushing edge as an extension of the cloud, and it's still anyone's game."

Breaking down Kubernetes edge computing competition

As in many areas of Kubernetes management, Red Hat's OpenShift platform has a slight lead over VMware Tanzu in rolling out features that support edge computing, such as support for single-node clusters.

Analysts at VMware's Explore conference this week asked VMware officials how the company will differentiate its edge computing stack amid strong competition from Red Hat and others; the answers boiled down to plans to compete on ease of use and integration with existing VMware products many companies already have, said Paul Nashawaty, an analyst at ESG.

"Once you're involved with VMware, it's really hard to unwind it and move to OpenShift," Nashawaty said.

VMware is also about to narrow the gap with OpenShift edge computing features. Red Hat OpenShift 4.9 supported single-node clusters in October 2021; Tanzu Kubernetes Grid 2.0 will add single-node support when it ships this October. OpenShift 4 began to support automated infrastructure updates for air-gapped, or offline environments in mid-2021; Tanzu Application Platform 1.3 will also support these features in October.

Both VMware and Red Hat offer broader packages of software in addition to Kubernetes to support edge computing. Tanzu Kubernetes Grid 2.0 is part of VMware's Edge Compute Stack 2, a bundle that includes VMware's ESXi VM hypervisor, vSAN software-defined storage and SD-WAN Service networking products. TKG 2.0 will support GPU pass-through, which routes data processing around software-defined infrastructure layers to boost performance in AI/ML edge computing environments; OpenShift also supports similar hardware acceleration techniques. VMware's vSphere 8 boosts support for real-time operating systems and non-x86 processor architectures; Red Hat Enterprise Linux (RHEL) 8 offered such integrations with RHEL for Real Time, released earlier this year.

Red Hat and VMware have recently begun to partner with mobile service providers and auto manufacturers to support edge computing applications. Dish Network has begun to roll out a 5G software-defined radio access network (RAN) using VMware software-defined networking tools; this week VMware also publicized a new partnership with NTT that will see the 5G carrier use VMware's edge network as a service products.

Meanwhile, Samsung uses Red Hat OpenShift in its 5G networks, as part of a years-long 5G partnership, and automaker GM kicked off a collaboration with Red Hat for an in-vehicle operating system in May. Not to be outdone, VMware announced this week that Audi has signed on to replace industrial PC hardware with the Edge Compute Stack.

The development of edge computing and Kubernetes so far have been intertwined, since containers lend themselves to resource-constrained edge devices more than virtual machines. Kubernetes automation, particularly through GitOps, can also help programmatically update many small clusters at the edge that can multiply beyond the capacity for manual management.

Thus, enterprises interested in edge computing also have a glut of choices among Kubernetes platforms other than VMware Tanzu and OpenShift that approach the market from various angles, from Kubernetes platforms to cloud services. In the corporate branch location realm, SUSE's Rancher was an early player in multi-cluster edge computing orchestration using Kubernetes and its stripped-down k3s distribution of Kubernetes optimized for edge environments. There are also open source options that support homegrown edge computing platforms such as the one Deustche Telekom developed using upstream Kubernetes and Flux GitOps for its 5G rollout last year.

Analysts spy edge computing opportunity with Broadcom

VMware's edge computing advancements have caught the attention of industry observers who see an obvious area where it can combine products with Broadcom Inc., which publicly disclosed its intent to acquire VMware for $61 billion in May. Despite its recent acquisitions of software companies in CA and Symantec, Broadcom's core business is in customized system-on-a-chip hardware used in mobile devices, including Apple's iPhone. It also makes chips used in data center switches by cloud providers including GCP and AWS.

AWS has made its own investments in silicon, but there are still areas where VMware and Broadcom could potentially combine forces to compete -- with an emphasis on potentially, according to Larry Carvalho, an independent analyst at RobustCloud.

"Together, the two companies have a chance to deliver unique solutions combining VMware software with Broadcom's ability to create built-for-purpose chips," Carvalho said. "However, the time this may take will be too long for [some] customers to wait."

One VMware customer said he agrees that Broadcom and VMware could produce unique edge computing products but hopes that the combined companies won't focus too much on the Kubernetes component.

"Virtualization is still a big deal, and there are still a lot of companies that prefer on-site stuff," said Brian Kirsch, an IT architect and instructor at Milwaukee Area Technical College. "A virtual stack on dedicated hardware, I think that would be great -- you could do that for building controllers and other HVAC-type systems. I think there's some really serious market there, but I don't know what's going to happen to the rest of the stack."

VMware executives mostly declined to comment on specific plans with Broadcom post-acquisition. CEO Raghu Raghuram acknowledged in a press conference at VMware Explore that there are obvious possible integration points for VMware's software and Broadcom's chips, but was careful to emphasize that VMware will not have any special preference for Broadcom edge computing products.

"Just like we did in the data center, we will work with not just Broadcom, but anybody that's producing chips, whether for networking, machine learning or core compute processing," Raghuram said.

Enterprise Strategy Group is a division of TechTarget.

Beth Pariseau, senior news writer at TechTarget, is an award-winning veteran of IT journalism. She can be reached at [email protected] or on Twitter @PariseauTT.

Dig Deeper on Containers and virtualization

Software Quality
App Architecture
Cloud Computing
Data Center