Alex Rousskov wrote:
>
> On Fri, 26 Mar 1999, Oskar Pearson wrote:
>
> > Assumptions: a cache digest is 200 000 bytes (a large peer may exceed
> > this quite easily, though). This is equivalent to 3333 ICP packets
> > (200000/60). If you retrieve the cache digest every 10 minutes
> > (600 seconds), you need to be sending 5 1/2 icp queries a second to
> > make cache digests worthwhile (3333/600). For ISP caches, cache
> > digests make sense. For ISP customers, you almost certainly want to
> > use ICP.
>
> Oscar,
>
> I am afraid this math (and the conclusion) is somewhat misleading
> even with the current [ugly] Digest implementation. Please correct me if I
> misinterpreted your example. Receiving a digest every ten minutes implies
> that you have 6 peers. Sending 6 ICP queries per second if you have 6
> peers means one query per second per peer. That is a very low rate for
> most configurations with large number of peers.
Umm. This is assuming only one request per-second, however, is it not? I
have a setup identical to this example, and each of the seven units is
sending ~42 ICP requests per second.
> Also, using as many as 6 peers with ICP often implies noticeable delays.
Well, yes. This is true.
> Finally, our initial experiments with "deltas" or "diffs" indicate that we
> can further [drastically] reduce bandwidth requirements for digests.
If it can be made smaller, that is (obviously) better. The current
implementation suits me fine, however.
D
Received on Fri Mar 26 1999 - 17:08:46 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:45:27 MST