ClusterHQ? More like Clusterfscked: Cluster bombs, says it came to market too early

Preempting cheap shots at his company's failure, CEO Mark Davis on Thursday announced that ClusterHQ is shutting down with a single, unassailably apt headline: "ClusterF***ed."

We couldn't have said it better ourselves at The Register. Our job is done, or nearly.

Given how rare it is that a technology company commits the heresy of journalism – saying something that deviates from the marketing message – we asked Davis to elaborate.

His valedictory blog post stops short of an explanation for declaring an immediate cessation of operations, offering only "a confluence of reasons" for the decision.

ClusterHQ offers container data management. Its software provides a way to manage stateful applications like databases in containers.

Containers started out as environments for stateless applications, which can be started, stopped, and restarted without the need to persist data. Stateful containers, capable of handling databases, represent the next logical step in the evolution of container technology.

"When Docker hit people's radar a few years ago, the idea was using containers as a way to build and deploy web-scale stateless apps," said Davis in a phone interview with The Register. "And the assumption at the time was the databases would be left on traditional VMs."

ClusterHQ, Davis said, came to market thinking the benefits of containers should be extended to databases.

As Davis points out in his post, container data management has begun attracting attention. Docker, for example, recently acquired Infinit to provide a data layer for its containers.

The Register asked Docker chief developer advocate Patrick Chanezon at the time whether this move toward a "batteries included" model – shipping with a default, albeit replaceable, data layer – might threaten other vendors focused on container-oriented storage. Chanezon insisted the Docker community welcomed a built-in storage option.

That's not to say Docker should bear the blame for ClusterHQ's demise. Rather, its interest in the container data layer can be seen as a bellwether for an idea that hasn't yet arrived.

Asked whether the company's failure could be attributed to a change in venture capital market, one that might herald an economic downturn, Davis dismissed the idea. Even so, he acknowledged, "the venture market not what it was when we did our series A."

ClusterHQ's problem, as Davis described it, was that the company came to market too early.

"The way we went through our decision process is looking at where the market is so far, and what the revenue opportunities are," said Davis. "While there's a big market to be built, it's off in the future. But not the near future."

There's broad interest in containers, said Davis, but much of the activity is experimentation rather than production workloads. "There are some exceptions," said Davis, "but the revenue for vendors who have to build this stuff is still relatively small."

Davis estimates that stateful containers lag the general container market by about two years.

ClusterHQ will go through a wind-down process, said Davis, adding that the company's open source software at least will continue to be available. That may not offer much holiday cheer to the company's 23 employees. ®