[Noisebridge-discuss] Distributed computing and storage (with some major caveats)
Sai Emrys
noisebridge at saizai.com
Tue Jun 23 22:53:28 UTC 2009
On Tue, Jun 23, 2009 at 7:04 AM, d p chang<pchang at macrovision.com> wrote:
>> All data should be replicated redundantly across nodes, so that it can
>> survive any node going down without loss of the data that that node
>> may have generated (or proxied).
>
> ok. i was mostly wondering if this was a 'simplification' in that nodes
> intentionally had 'partitions' of the complete data set.
No; if it were simple partitions, then it'd be highly vulnerable to
nodes going down.
It might be possible that some RAID variant would work acceptably
well, but I suspect not. I know there are systems that do this (eg
dhash?), but I don't know how they work.
- Sai
More information about the Noisebridge-discuss
mailing list