thundercloud computing

the closed world of web 2.0

The end of shrink-wrapped software is in sight. This is also a threat to the social contract of the open source community, which is constructed around the distribution of software. The web 2.0 world doesn't distribute software anymore, it just offers a browser view to an application that actually runs in a corporate server farm. To put it bluntly, Google can (and does) take GPL software and effectively make it proprietary.

On top of that, as a user you not only lose your software freedoms, you also lose control over your data. All those web 2.0 sites keep your data on their servers, in their formats. So it actually ain't your data anymore, is it? Dataportability.org aims to address this problem. The effort is laudable but hasn't produced any technology yet.

open cloud computing

The Hadoop project features an open-source cloud computing stack that's based on Google's MapReduce algorithms and GFS distributed filesystem.

An open source cloud computing capability is a necessary, but not a sufficient condition, for software freedom in a web 2.0 world. We need software freedom at the application layer, not just at the resource layer that cloud operating systems are concerned with.

think big

We need to think very hard about ways to construct distributed open source applications, mash-ups if you will, that

1. are able to compete effectively with closed alternatives in the same web 2.0 application space
2. cannot be coopted and turned proprietary
3. are able to guarantee security and trust in distributed operations

Centralized, proprietary web2.0 services make us unfree. We need, not just another social network, but another social networking stack. A distributed, decentralized cloud capability with the open-sharing ethic hardcoded into it's technical genome.

We can work from the leaves, from the gigantic potential of underutilized computing resources on the edge, the idle cycles in your laptop right now. We don't need to compete in the center, with multi-billion hydropowered data centers. We can assemble a computing swarm with equivalent power from end-user devices.

We just need to be smart, and come up with an open source cloud algorithm that puts control in the hands of the end-user (you!). A resource sharing protocol that resists intrusion, promotes sharing and scales like crazy. A crossover between Bittorrent, Xen and Hadoop.

Spammers create cloud-scale botnets; of course we can create cloud-scale free software. The nice thing is: it will happen. History has shown, that somebody, somewhere, will stand up and make it so.