One week ago, a one hundred and seven year old technology company bet its future, at least in part, on an open source project that turned four this past June. It shouldn’t come as a total surprise, therefore, that a twenty year old six hundred pound gorilla of virtualization paid a premium for one of the best regarded collections of talent of that same open source project, the fact that containers are disruptive to classic virtualization notwithstanding.
But just because it shouldn’t come as a surprise in a rapidly consolidating and Kubernetes obsessed market doesn’t mean the rationale or the implications are immediately obvious. To explore the questions of why VMware paid an undisclosed but reportedly substantial sum for Heptio, then, let’s examine what it means for the market, for Heptio and for VMware in order.
For the Market
If there was any question in the wake of IBM’s $34 billion acquisition of Red Hat and its Kubernetes-based OpenShift offering that it’s Kubernetes’ world and we’re all just living in it, those should be over. This acquisition means that VMware, like IBM, has placed a bet not just on containers as one path but Kubernetes specifically – in a fashion similar to its spinout Pivotal via PKE. In this, they join a number of companies that have bought into that burgeoning market:
- Apprenda (Kismatic)
- Bitnami (Skippbox)
- Microsoft (Deis)
- Oracle (Wercker)
- Red Hat (CoreOS)
- New Relic (CoScale)
- IBM (Red Hat)
Perhaps more importantly, the Kubernetes standard is shared by the Big Four hyperscale cloud providers, in Alibaba, Amazon, Microsoft, and project founder Google, each of which has Kubernetes-based offerings to sell.
From a macro perspective, then, whether a business is working with the largest of the large cloud providers, tiny nimble startups or more traditionally focused enterprise infrastructure suppliers, it is likely to be encountering Kubernetes – whether it likes it or not.
What this means, then, is that containers broadly and Kubernetes specifically is becoming the next Java. Not in the technical sense, obviously; one is a programming language and runtime, the other is a container orchestrator. But from a workload portability standpoint, their functional jobs to be done are on a converging path.
This doesn’t necessitate Kubernetes, notably; as one of the project’s advocates Jessie Frazzelle has written, there are many alternatives. And conversely, as with every technology phenomenon, my colleague has noted there are a host of enterprises that are rushing to deploy Kubernetes because it’s popular, without knowing or having thought through precisely what they need it for.
For better or for worse, however, technology is a fashion industry, and Kubernetes is for the moment – and likely for the future, given the level of investments and commitments – haute couture. It will be interesting, therefore, to see which remaining Kubernetes talent pools are acquired, by whom and at what price. It may yet be that Red Hat’s somewhat surprising valuation of CoreOS proves to be a bargain.
While every founder tends to have visions of large, independent entities, this acquisition is a near optimal outcome for Heptio. From an analyst’s perspective, Heptio has always been an interesting startup to follow. The assumption for many in the industry following the initial founding was that they would pursue a similar product course to other commercial open source organizations, which is to say that it would become something very similar to a traditional product company. Between being founded by two of the project’s founders and the quality of the technical talent they were able to recruit and surround themselves with, the opportunity for differentiation would have been there, even if the company’s primary in market competitors were much larger and infinitely better capitalized.
Instead, the company chose a unique path of not quite product company, and not quite full service company, but borrowing elements of both. This fit the market need in many cases, but posed significant challenges from a marketing and messaging standpoint, as the market understands product companies and service companies but is less comfortable with descriptions that don’t entirely fit into either bucket.
Throw in the challenges of competing with ever larger entities, as Kubernetes-driven consolidation continued apace, and it’s clear why VMware would be an attractive home for the startup given a substantial enough payout and ability to retain the necessary talent. VMware’s immense account reach would be an immediate stimulant for the sales pipeline of a smaller, more resource-strapped startup, and even better, said accounts would likely be receptive to the messaging. Heptio, after all, has what many VMware customers want: container capabilities. Container capabilities, importantly, that VMware will via Heptio be able to weld more tightly to its base virtualization offerings to at once enhance the appeal of their existing product set and – in a perfect world – jujitsu the transition to containers from threat to opportunity.
VMware is a company with immense weight and inertia within the datacenter, but it is also a company that would benefit from both a platform more relevant to developers and a team capable of attracting them. Hence, this acquisition and, presumably, the valuation.
A decade or more ago, developers were attracted to virtualization and to VMware because it solved problems from dependency management to testing. Over the years, as products arrived from Vagrant to Docker, developers found lighter weight solutions to address some of the same pain points they once attacked with VMware offerings. And importantly, many of these options were open source, and as a consequence, free.
None of which concerned VMware particularly, because as its early developer attention waned its popularity within the operations side of the house – and central IT in particular – boomed. Which has been good for the company generally as central IT has historically been less concerned both about software being open source and being free than developers, which has led to VMware becoming an enterprise datacenter standard which in turn led to its current $60.7B market cap – a valuation roughly double that of Red Hat’s, for context.
In recent years, however, the incredible popularity of containers amongst developers – RedMonk has never seen any technology grow more quickly – created demand for ways to deploy and manage them in production. While Kubernetes developers then accurately point out that the software is at least as much if not more so optimized for Day Two or Three operations – an operator’s concern, then, rather than a developer’s – it is improbable that Kubernetes would be the phenomenon it has become absent the catalyst that is developer’s insatiable appetite for containers.
What this means in practical terms then is that VMware’s classic virtualization business had to view containers as a potential disruption at a minimum. To date, actual container traction in production has lagged not only the hype around the technology but virtual machine based alternatives substantially. But if developers loved containers enough that operations was forced to find a way to run and manage them in production, at some point this presumably would begin to impact adoption of virtual machine based technology and thus VMware. The company therefore had two approaches: fight the tide, or embrace and extend it.
With this acquisition, Kubernetes becomes not just the de facto container standard for VMware if it wasn’t already, but the engineering focus to lead with both internally and externally. Internally, the hope is that Kubernetes and the remarkable developer team Heptio had assembled, can bring some of that project’s momentum and drive to a talented but generally conservative, proprietary and IT-centric VMware engineering staff. Externally, meanwhile, the company stands to acquire one of the better Kubernetes developer relations capabilities in the industry.
As with all acquisitions, this one will depend on execution and there are no guarantees of success. But on paper, at least, this seems to be a logical and potentially more than the sum of its parts combination.
Disclosure: VMware is a RedMonk client, as are Amazon, Bitnami, Google, IBM, Microsoft, New Relic, Oracle, Pivotal, Red Hat. Heptio and Alibaba are not RedMonk clients.
Credit: Information on acquisitions was compiled by my colleague, Rachel Stephens.