Who controls Vert.x: Red Hat, VMware, or neither?

Vert.x is an asynchronous, event-driven open source framework running on the JVM. It supports the most popular Web programming languages, including Java, JavaScript, Groovy, Ruby, and Python. It’s getting lots of attention, though not necessarily for the right reasons.

A developer by the name of Tim Fox, who worked at VMware until recently, led the Vert.x project — before VMware’s lawyers forced him to hand over the Vert.x domain, blog, and Google Group. Ironically, the publicity around this action has helped introduce a great technology with an important future to the world. The dustup also illustrates how corporate politics works in the age of open source: As corporate giants grasp for control, community foresight ensures the open development of innovative technology carries on.

Vert.x under the hood

Think of Vert.x as Node.js written for the JVM, a simplifying technology for advanced solutions in the cloud. It has a simple concurrency model that enables developers to write single-threaded applications, but provides a nonblocking event model and distributed event bus so that sophisticated, scalable, distributed applications can be developed. It potentially has access to the full range of Java class libraries and can be used with a full range of enterprise applications if programmed in a suitable language.

Vert.x was largely developed by its project lead, Tim Fox, who worked for VMware’s SpringSource unit — until he was hired by Red Hat last week. While Vert.x uses existing ideas from Netty (recently spun out from JBoss as an independent community), Vert.x is a complete platform for writing asynchronous applications and is clearly a very powerful tool for the current generation of Web applications. As a consequence, a growing crowd of developers has embraced it; indeed, Nate McCall of Apigee noted they had “recently staked a lot on vert.x” and another list member suggested there were “millions of dollars” of development at stake.

The project looks safe enough from a developer perspective. It’s hosted on GitHub and licensed under the Apache license, so every community member has equal access and is free to fork. All the same, VMware had been accumulating copyright assignments, and the company’s response to Red Hat hiring Tim Fox was, according to a post he made to the Vert.x community, to call in all his community roles as if they owned them, using legal threats in the end. This move immediately raised concerns, even provoking calls for a fork and comparisons to the Hudson/Jenkins split.

Red Hat politics

Red Hat declined my invitations to comment on the situation, so I’m left to analyze the evidence available in the community to try to understand what’s going on. This is clearly an area Red Hat’s JBoss unit cares about, and Mark Little — the Red Hat executive who commented officially on the mailing list — is its CTO. It seems Tim has been hired competitively from VMware to work on Vert.x full time. Further, Tim’s original message was approved for release, so the challenge to VMware was intentional. While it’s possible this is a by-product of the well-known rivalry between JBoss and SpringSource, dating from before either was purchased, it seems more likely this is part of Red Hat’s strategy.

What is that strategy? I can only guess, but the various comments from VMware employees in the mailing list thread suggest more is going on behind the scenes. Red Hat’s public statements seem tuned both to move Vert.x out of VMware’s control to an open source foundation and to make a community-friendly public response by VMware difficult. Whatever the plan, Red Hat and VMware have called for “input” from the community about the future of the project — a move welcomed by most voices in the conversation — and while they are clearly engaged in private negotiations about the future of Vert.x, an open discussion has started.

What would be best for the community? An outcome that allows both Red Hat and VMware to continue their investment in Vert.x would be preferred. That may involve moving the project into an established foundation like the Apache Software Foundation or the Eclipse Foundation — indeed, Apache president Jim Jagielski and Eclipse executive director Mike Milinkovic have already dropped by to offer help — but that solution may be a little process-heavy. A looser approach may be more appropriate, using independently hosted code and wiki sites and depositing any shared assets in a nonprofit like SPI. Whatever happens, the existing Apache open source license used by Vert.x already ensures every participant has the freedoms they need.

Power play There’s a community lesson to be learned here. If you’re going to depend on a technology, the license and the governance aren’t dull bureaucracy — they matter. In this case, the use of a well-understood open source license guaranteed the liberties needed by the community, but the absence of community-centric governance risked disruption. Projects anchored on independent, nonprofit organizations are protection against this — hence their growing popularity.

There’s also a lesson in the new corporate politics. In an age of open source, it’s hard to acquire a technology. When Oracle acquired Sun, the open source technologies didn’t automatically come with it despite copyright ownership. Almost all the MySQL developers went elsewhere, especially to work on MariaDB. Both the technology and the existing business around identity software went elsewhere with the key staff. Oracle may have thought it could control Hudson and OpenOffice.org, neither of which it considered strategic, but both of which it thought it could manipulate for marginal political benefits. In both cases, the core of the community moved away — to Jenkins and LibreOffice — and neutralized the power play.

Acquiring a company is expensive and has a poor track record of success even with proprietary technologies. For open source, trying to acquire a technology with a real community is hard, unless you have a strong understanding of both developers and community dynamics. Control doesn’t win you much on its own; you need influence.

The most important step is to recruit key community members and gain crucial influence over the project. By poaching Tim Fox from VMware, Red Hat has made a chess move derived from extensive experience of open source. It’s gained control over future development of Vert.x, triggered a move to independent governance, and negatively framed VMware. This is the 21st-century equivalent of a hostile takeover, as played by experts. I don’t think the game is over yet.

This article, “Who controls Vert.x: Red Hat, VMware, or neither?,” was originally published at InfoWorld.com. Read more of the Open Sources blog and follow the latest developments in open source at InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.