2 ms·
I think the model would have to be to develop an open source kit that allows standardized pluggable distributed hosting in "a box". I wouldn't expect a signifi
by mseebach 18y ago
I think the model would have to be to develop an open source kit that allows standardized pluggable distributed hosting in "a box".
I wouldn't expect a significant market for simple non-dynamic hosting - the solution would have to accomodate blogs, forums and other dynamic content-apps.
Any provider of such hosting would just be provided with a config-file, and then start to pull the content, sync the DB etc. The kit should be designed to expect nodes to drop of randomly, and be self-healing (i.e. "elect" a new master-node, or be purely peer-to-peer).
Maybe the kit could be modelled on top of Google App Engine.
One single-point-of-failure would be DNS though. If your domain is .com, and Verisign pulls it, you could be hosted on the moon, and still disappear.
- jbert 18y ago> If your domain is .com, and Verisign pulls it, you could be hosted on the moon, and still disappear. Ouch. I forgot about that. I guess one approach would be to have the same name under several registries, trying to get decent coverage across juridictions where the physical nameservers are located. You could encourage people to choose a random (supported) top-level part (.com, .fr, .se, .co.uk, etc) in links, so a single DNS takedown would only hit a portion of links (which could then still be hand-edited to a working link). Network resilience and peering would also be an issue. A country can always filter you out I guess, a la Great Firewall. The sync issues would be interesting. Not sure what the people using the hosting see (a db? a filesystem?). Do-able, tho. (Off the top of my head, I'd go for a DVCS-like model, where it's fully peer-to-peer, automerge when we can, punt to a human when we can't).