[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [GNUnet-developers] GNUnet replication mechanism

From: Christian Grothoff
Subject: Re: [GNUnet-developers] GNUnet replication mechanism
Date: Wed, 2 Apr 2003 12:18:26 -0500
User-agent: KMail/1.4.3

Hash: SHA1

On Wednesday 02 April 2003 03:01 am, eric haumant wrote:
> Is it possible to make GNUnet send more noise to other peers. It would be
> interesting to add a key in gnunet.conf to set the noise padding rate to
> make gnunet replicate more random data.

I'm not sure if this is a good idea since it will also increase bandwidth 
consumption and noise is generally not a good thing. If you want to play with 
it, one good way may be to make the value of TTL_DECREMENT 
(afs/protocol/messages.h) smaller, say 200 * cronMILLIS instead of 5 
cronSECONDS. That would decrease the average delay that GNUnet applies to AFS 
messages and thus prevent GNUnet from batching more messages together and 
thus incrementing the likelyhood of sending noise.

This will only work for peers that are actively routing AFS messages and not 
for completely idle peers, though. Another possibility is to add a cron-job 
that gets some random content and sends it out, that should not be too hard 

> I haven't get yet throught the core source code, can you tell me where
> (which file and which function) is gnunet generating random noise messages
> ? It's probably a call to sendBuffer or appendToBuffer in connection.c but
> I can't find the real call. There isn't any log there.

GNUnet is not generating random noise messages, all it does is padd messages 
with random noise to make them uniformly sized. This is done in sendBuffer. 
The assumption is, that a P2P network typically has enough background noise 
(or enough activity in general) to avoid the need for additional artificial 
noise generation beyond the padding. 

Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see


reply via email to

[Prev in Thread] Current Thread [Next in Thread]