Hope you don't mind me taking a swing at these.

> 1) What causes specialisation in nodes? Is the
> "specialisation" area "picked" 
> by a node using some method, e.g. most frequently
> used parts of key space, as 
> perceived by the node through incoming requests, or
> is it merely amplified 
> from random artefact clustering around certain parts
> of the key space?
They don't want a node to be able to pick it's
specialization, so it's "merely amplified from random
artefact".  I don't believe this elliminates the risk
of an evil node choosing it's specialization.

> 3) Is probabalistic caching is used to amplify
> specialisation? From previous 
> discussion, I seem to remember it being said that a
> node will cache 
> everything until the space runs out, at which point
> it will start to drop 
> data using "probabalistic caching". How is this
> traded off between 
> maintaining specialisation in parts of key space and
> caching the "most 
> popular" data? The two can contradict each other,
> e.g. the least 
> frequently/recently used file can be in the key
> space of strong, narrow 
> specialisation.
Try to think of the caching in two pieces A) stuff
get's in B) how stuff get's thrown out.

A) The probablistic caching doesn't kick in for
newbies; when the node was 75% (not sure exact number)
full then only some of the objects are cached.

B) As far as I know they still use simple least
recently used (LRU); when the A causes the cache to
overflow the oldest thing gets chunked.

The delayed kicking in of probablistic caching
explains this change in rate of datastore growth you
noticed.

There sometimes seems to be a bit of circular logic
going on:
A specialized node will get data for its keyspace.
A node will become specialized if it has data for the
keyspace.
or
The routing works because the caching works.
The caching works because the routing works.

There may be some neat tweaking possible where things
within your specialization are kept a bit longer than
other things, or get a higher chance of getting
cached.  It's not as much a liablity to have data in
my store which is in my specialization, it could have
come from anywhere.  It's also more important that the
nodes specialized in the area have the data than
someone else.  They insure the data is reachable, the
others just help with performance. 

__________________________________________________________________

Gesendet von Yahoo! Mail - http://mail.yahoo.de
Logos und Klingelt�ne f�rs Handy bei http://sms.yahoo.de
_______________________________________________
Devl mailing list
[EMAIL PROTECTED]
http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to