VMware By Broadcom - What Happens Next?
A lot of my discussions both internally and externally over the past several weeks have been on the upcoming Cloud Service Provider changes that are happening in the VMware space following the Broadcom transition. While there’s still a lot that is up in the air, I think there are a few points that will converge over time as Broadcom shows their new strategy.
More Focus on Development, Less on Operations
VMware had a long-term goal to improve revenue and profitability through increased offerings of their solutions “as-a-Service”, wherein they would be able to increase wallet share through selling offerings that included both the software and platform that it runs on, as well as reducing costs through economies of scale in deploying these platforms. Conversely, Broadcom seem to have largely taken the view that their position is to create software, not run it, as as such are actively divesting themselves of these platforms and the associated overheads. Given VMware’s background is in software development, not operations of platforms, this is probably a more viable concept, which Broadcom have then extended by pushing the onus of providing support towards the service provider for most incidents, further reducing the support overheads needed. This cost reduction will free additional budget to focus on additional development to improve products and margins for the company.
Reducing the Solution Stack
One of the big shocks to many was the sudden announcement of several product SKU’s being discontinued effective immediately, which was widely misinterpreted by several outlets. While the variety provided by having these options allowed for near-infinite customisations for customers to consume what they wanted, it also created a development headache in ensuring that all the products and all the ways they interacted were supported and supportable. Through reducing the stack to two main flavours, VCF and VVF, there is a more streamlined approach to what is supported by VMware, reducing the wasted development cycles in ensuring backwards compatibility and instead working on forward-looking interoperability for the next release of the VxF bundle. Then to extend the push for customers to adopt a full-stack approach, by lowering support for those who aren’t utilising the full stack, there is an implicit pressure for customers to deploy a uniform stack across all implementations. This has the flow-on effect of reducing not only the variances in what is deployed on-site, but also in what the support team from partners or from VMware need to be aware of when responding to incidents.
Homogeneity
As we track towards a set of unified solution stacks, with support for those stacks being provided by partners instead of Broadcom in most circumstances, and a single means of transacting (as a subscription, using a universal metric), we can begin to see each Cloud Service Provider as one of many individual nodes across a wider VMware Cloud, including those operated by Hyperscalers, instead of a set of individual environments. This will help Broadcom in two ways, one by reducing the impact of a given partner in the commercial relationship, as they become easier to transition between environments given the technology consistency across the platforms. The second is that again because of the technology consistency, the cost of providing that service is pushed down by natural competition, while they have no inherent need to reduce their own prices given their solution’s position in the environment. In the same way we don’t tend to care about whether a given watt of electricity comes from solar or hydro, it will not matter whether a VM is hosted at provider A or B, save from needs around efficiency between the source and destination.
Bringing the three together, we end up with a strategy that looks to reduce the level of variation that exists in the market across providers, as well as pushing some of the onus of running and supporting the platforms to those running it, with the net goal being that the aggregate capabilities of VMware-powered cloud systems act as a “dis-aggregated hyperscaler” of sorts, similar to the original vision of the vCloud Air Network from years ago. This model would ideally allow Broadcom to improve their ROI by focusing on a small number of customers to support, as well as a narrower and deeper product focus, allow partners to focus on cost efficiency and economies of scale for their customers, and for customers provide the opportunity to have a sense of mobility when moving across different providers, due to the technological similarities that exist. While not a perfect solution, and one that will undoubtedly ruffle many feathers, there is a method that exists in the Broadcom approach that has a viable chance of success, which is something that could not be said for most of VMware’s approaches to market over the past decade or so.
