I think the real question is how failures are correlated and what the recovery expectations are.
If you look at a company with staff, and consider disks independent, then many faliures are just a disk. If the motherboard fails, you'd have a spare and move the disks. But there is also a fire/flood in the datacenter. In a cypherpunk-style grid, you have many people and any of them might go away. So we really need to know and have expressed about administrative control and physical correlation, and then a way to talk about what we care about and what might happen, all of this driving share placement. I know, ENOPATCH.
pgptoNlreqbln.pgp
Description: PGP signature
_______________________________________________ tahoe-dev mailing list tahoe-dev@tahoe-lafs.org http://tahoe-lafs.org/cgi-bin/mailman/listinfo/tahoe-dev
