Last week's OSCON conference served to remind us that open source software is setting the pace. We've come a very long way from the old saw that "open source doesn't innovate." Instead, you might ask: Is innovation in enterprise software happening anywhere else other than in open source land?
Hadoop is at the center of the big data trend. OpenStack has the momentum in private cloud. Open source frameworks and IDEs absolutely dominate app dev, while all the leading NoSQL databases are open source. Do I need to mention that Android now powers more smartphones than any other mobile OS? Plus, Microsoft and Salesforce excepted, you'd be hard-pressed to find a cloud provider that uses anything but open source software to deliver its service.
IBM's public embrace of Cloud Foundry at OSCON provides a telling example of open source's pole position. As with OpenStack, IBM is providing code contributions, but the Cloud Foundry community will steer development. IBM, which loves to use open source software for its consulting engagements, reaps the benefits.
Jockeying for position
The Cloud Foundry example also shows that open source has become a stage where the shadowplay of big vendor politics ensues. Cloud Foundry is a core technology of Pivotal, spun out from VMware in April. From the beginning, the downloadable version of the Cloud Foundry PaaS (still in beta) was open source -- VMware built it and in 2011 made it available under an Apache 2 license. IBM's announcement that it will incorporate Cloud Foundry in its cloud platform and join the advisory board represents a strategic alliance with Pivotal.
How might that affect IBM's alliance with Red Hat, whose OpenShift platform competes directly with Cloud Foundry? I can imagine there are some unhappy campers at Red Hat right now. But IBM is smart -- as with OpenStack, it has very likely backed the winning horse. Cloud Foundry is one of the few PaaS plays that manifests itself as both a cloud service and locally installable version that will have special appeal to enterprises. Not to mention that in the recent InfoWorld comparison "Which freaking PaaS should I use?" Cloud Foundry came out on top.
But don't feel too sorry for Red Hat. It has become the greatest contributor of code to OpenStack, and the company's KVM hypervisor underlies the vast majority of OpenStack implementations. That has raised fears in some circles that the open source company might already have a lock on the OpenStack project, although Rackspace got the ball rolling. RHEL OpenStack Platform may one day grab hold of the data center the way Red Hat Enterprise Linux itself has dominated.
On the Hadoop front, it's clearly a pitched battle between Cloudera and Hortonworks. The former has achieved the greatest commercial success with enterprises and employs as chief architect Doug Cutting, the brilliant technologist who started Hadoop while working for Yahoo. But Hortonworks, the Yahoo spin-off and true open source steward of the Apache 2 Hadoop project, has steadily gained more corporate customers -- in part through partnerships with Teradata, a traditional BI player, as well as Microsoft.
In the NoSQL realm, we have true examples of pure-play open source leadership. The open source favorites Cassandra, Couchbase, and MongoDB have emerged on top. 10gen founder Dwight Merriman started the MongoDB project because he saw the need for a new type of database for Web-scale projects while working at DoubleClick. CouchDB, the open source project that provided the core technology for Couchbase, was created by leaders of the BSD-licensed memcached project. The technology behind Cassandra was originally developed at Facebook, but the highly active Cassandra community has taken it further, while open source vendor DataStax has packaged Cassandra with Apache Hadoop and a pile of enterprise security features in its Enterprise version.
Community development
Every open source project has its own dynamic. When big vendors get involved -- as with OpenStack or the new OpenDaylight project -- you know the political infighting will be intense. But interesting things happen when a winner emerges. For example, developers often say that they love the convenience of PaaS, until they discover a missing or awkwardly implemented feature that makes their lives hell with no obvious workaround. If Cloud Foundry emerges as a dominant platform, you'll have a huge community making contributions to offer a more complete, viable solution to make that less likely to happen.
Instead of hashing out the future in endless standards committee meetings, we're hammering it out in open source bits.
Clearly, some projects are authentically grassroots -- Mozilla Firefox, still listed as the most popular project on Ohloh, provides a shining example. Then there are grassroots projects created in reaction to bad vendor behavior, such as LibreOffice or MariaDB (or Linux, for that matter). Others, such as Google Android or Chromium, are firmly under a single corporation's wing. Ultimately, it's hard to get away from the fact that in high-profile projects, the most active committers tend to be employed by a commercial vendor with a specific agenda. Nonetheless, continuous involvement by a broader community yields better software.
This year, for the first time, respondents to the annual Future of Open Source Survey chose "better software quality" as the No. 1 reason for adopting open source. Sounds like the word is getting around.
This article, "Open source races to the top," originally appeared at InfoWorld.com. Read more of Eric Knorr's Modernizing IT blog. And for the latest business technology news, follow InfoWorld on Twitter.