Does Cloud Computing Change the Open Source Rules?

by Mike Gunderloy - Aug. 01, 2008Comments (0)

There's been a lot of talk about cloud computing lately, including some excellent crystal-ball reading from our parent blog GigaOM. But it's an essay from the ever-interesting Tim O'Reilly that brings together the cloud and the future of open source - and some of his conclusions may distress those who are firmly convinced that open source licenses are the only way forward.

O'Reilly starts off by identifying four things he sees as key drivers of the success of open source. It's not intended to be an exhaustive list, but it's a good one:

Licenses that permit and encourage redistribution, modification, and even forking;
An architecture that enables programs to be used as components where-ever possible, and extended rather than replaced to provide new functionality;
Low barriers for new users to try the software;
Low barriers for developers to build new applications and share them with the world.

Starting from this point, he considers the rapidly-changing landscape of open APIs, open data, portable virtual machines, and licensing issues. One of his key conclusions:

if you care about open source for the cloud, build on services that are designed to be federated rather than centralized. Architecture trumps licensing any time.

Those who are strictly focused on trying to get the infrastructure licensed as open source are probably missing the big picture. If one of the big issues of the open source movement as originally conceived was to prevent vendor lock-in, we need to recognize that lock-in can come through owning the data or owning the repository just as much as it can come from owning the software. If you depend on a single vendor to host everything, you're at that vendor's mercy.

Federation based on open standards provides a potential end-run around this problem. If you build your application to store multiple copies of data in a federated cloud, and the cloud members communicate based on open protocols, then in a sense you're vendor-proofing yourself. At that point you depend on your technical chops to spin up more instances that can talk to your existing instances even if their internals are different.

Of course, things aren't quite this simple: at the end of the day, adding more members to a federated cloud takes money (though we have examples going back at least as far as Fidonet of that being done in a decentralized, volunteer-funded way). As long as they're willing to engage in loss-leader pricing, the early cloud incumbents like Amazon and Google (with IBM, HP, and Microsoft waiting in the wings) are likely to have an advantage. But that doesn't mean that open-cloud proponents are out of luck: we should be able to flourish beside proprietary clouds, just as open source software grew up alongside its proprietary brethren.