Skip to Content.
Sympa Menu

bluesky - Re: Content based addressing

bluesky AT lists.ibiblio.org

Subject: Global-Scale Distributed Storage Systems

List archive

Chronological Thread  
  • From: dmolnar <dmolnar AT hcs.harvard.edu>
  • To: Global-Scale Distributed Storage Systems <bluesky AT ibiblio.org>
  • Subject: Re: Content based addressing
  • Date: Thu, 15 Feb 2001 22:24:42 -0500 (EST)
  • Resent-date: Thu, 15 Feb 2001 22:26:02 -0500 (EST)
  • Resent-from: dmolnar <dmolnar AT hcs.harvard.edu>
  • Resent-to: bluesky AT franklin.oit.unc.edu




On Thu, 15 Feb 2001, Wei Dai wrote:

> The second problem is a flat landscape. Suppose you insert a document
> with a key near the key of a very popular document that is cached
> everywhere. The insert request and subsequent data requests for this
> document will then be routed randomly and probably no one will be able to
> find it.

I know of no provisions in the Freenet protocol for dealing with either
problem. I haven't studied the protocol much, though. The idea I want to
point out is the one used by the OceanStore project at Berkeley
(www.oceanstore.cs.berkeley.edu). Queries in OceanStore have two kinds - a
quick, prone to error query and a much slower, but nearly guaranteed to
work query. The quick query is a randomized hill-climbing scheme.

The claim seems to be that the slow queries always succeed, while the
quick queries almost always succeed. Then if pathologies of the kind you
describe are rare in practice, the system runs quickly.

While it's still really, really important to characterize which search
keys lead to pathology in a "quick" scheme and avoid them, maybe having
a backup scheme makes it less critical in practice?

Thanks,
-David






Archive powered by MHonArc 2.6.24.

Top of Page