[FoRK] Our dispersed, not distributed, computing future

Adam L Beberg beberg at mithral.com
Mon Nov 9 13:28:23 PST 2009


Jeff Bone wrote on 11/7/2009 2:15 PM:
> The implications are obvious:  replication rather than immediate remote 
> access, explicit remote interaction rather than implicit, explicitly 
> latency-aware software with appropriate failure modes, loose / eventual 
> consistency protocols, caching (and timeouts) everywhere, loose time 
> protocols that preserve causality *as much as is physically possible* 
> (and no more --- cf. Lamport) --- and generally store-and-forward / 
> queueing-friendly architectures rather than synchronous, blocking 
> architectures.

Yea, and? We knew all that ages ago. How else would you even design the 
networking? I can't remember a time I didn't design systems like that. I 
remember seeing his DTN work and thinking it was cool, due to the whole 
space angle no doubt, until I realized that was just now I was always 
taught to do things.

But right now it's deeply unpopular, if not a dark art. There is just so 
much more money to be made with centralized machines and data hostage 
business models, and so few developers that haven't always had ~infinite 
memory/disk/cpu/bandwidth on their desktop. A model that's ripe for some 
serious disruption of course ;)

The funny part is Vint and the usual CS-ancestral crew knew all this 
back then too, that's just how you did things when the network being 
online was the exception, just look at how email works. The part that 
makes me wonder is what made him bring it up now, esp with the obvious 
conflict to his employer. The emeriti effect kicking in I suppose.

-- 
Adam L. Beberg
http://www.mithral.com/~beberg/


More information about the FoRK mailing list