Have Skills, Will Travel Charles Armour

Cloud vs. Mesh (and Google Wave)

Lately I’ve been reading a lot of news about fairly high profile datacenters and cloud platforms suffering failures, even Google’s App Engine.  It made me wonder if it’s time to re-visit the idea of a mesh cloud, which for some reason seems to be stymied in “impossible” land.  To be precise, we need something to get App Engine (or Google Docs, etc) off their cloud and onto an unbreakable distributed system.

Just the term “mesh cloud” would probably make a venture capitalist drool with anticipation, but the basic technology required for such a system is just a distributed data store.  If you haven’t yet heard of Wuala, check out their site and service.  I think they’re on the forefront of a viable technology for a mesh cloud computing, and boy do I wish they’d hire me (yes, I applied)!  Given Wuala’s decentralized and distributed data store (with proper servers as well as a mesh backing your data), the issue becomes one of bandwidth.

Since bandwidth-constrained applications do poorly in meshes, my thoughts centered on collaboration and communication software like Google Wave rather than data-intense applications like video encoding or real time simulation.  It’s really unfortunate that Google Wave requires a centralized server for locking and updates, as that breaks the technology for distributed platforms that lack locking semantics.  Of course, that may not be the case as information about Wave is hard to come by (shame on Google)!

So the real holy grail would be a cloud-based wave-like application with free, perpetual persistence.  I’ve put some thought into such platforms and would really like to dig in, but to get started I’d need to implement the very basics for a secure platform – a trust hierarchy.  While I could roll my own PKI system, I’m looking for something more like OpenID with a PKI component.  Is there any such beast?  Should there be?