On Fri, Mar 28, 2003 at 04:08:31PM -0800, Ian Clarke wrote: > We need to get probabilistic caching in the code ASAP - the current > level of load most nodes are getting is totally disrupting the > clustering effect from what I have seen in my node's diagnostics. > Freenet simply won't work properly without this clustering. > > I propose that we have a fall-off parameter which adjusts how the cache > probabilty "falls off" as the reply gets further from its source. This > can be configurable. We can put this functionality in unstable, and > encourage people to experiment with different parameters. We should see > increased clustering on those nodes irrespective of whether the rest of > the network contains that code. > > I propose making this the highest priority at this time. > > Ian.
After a conversation with oskar, we have fleshed out the following
proposal:
New config parameters:
Full Name: Target cache nodes per request per reset
Default: 5
We plug this into the formula:
<hobx> if we take N steps until reset on average, then we will have (1 -
p^N)/(1 - p) caches per reset.
To get the base p to put into the following equasion:
P(Cache data) = p ^ hopsSinceReset
hopsSinceReset is a new property on DataReply and InsertReply messages,
which is reset to 0 when the datasource is reset, and set to 1 on the
node that produced the data from cache in a request (the current
behaviour is for a node that serves a request from the datastore to send
the ref of the node it got it from, unless it is no longer there, it got
it from a transient or locally, in which case it resets it to itself; it
probably also does the usual probabilistic reset. Obviously if the
source is reset even on the origin node, we will set hopsSinceReset=0).
--
Matthew Toseland
[EMAIL PROTECTED]/[EMAIL PROTECTED]
Full time freenet hacker.
http://freenetproject.org/
Freenet Distribution Node (temporary) at
http://80-192-4-36.cable.ubr09.na.blueyonder.co.uk:8889/fl4hFW7uLxk/
ICTHUS.
pgp00000.pgp
Description: PGP signature
