As long as everyone keeps inserting everything onto all nodes, the Freenet
search algorithm won't work.  There will be no way to follow a path to a
particular node where the data is stored, if all nodes have approximately
the same data.  This causes requests to fail, so people try to fix it
by inserting onto more nodes, thereby making the problem worse.

Hal

_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl
>From - Fri Apr 20 11:46:52 2001
Return-Path: <devl-admin at freenetproject.org>
Received: from hawk.freenetproject.org (postfix@[4.18.42.11])
        by funky.danky.com (8.9.3/8.8.7) with ESMTP id MAA24932
        for <danello at danky.com>; Thu, 19 Apr 2001 12:35:54 -0400
Received: from hawk.freenetproject.org (localhost [127.0.0.1])
        by hawk.freenetproject.org (Postfix) with ESMTP
        id F33255818D; Thu, 19 Apr 2001 09:17:07 -0700 (PDT)
Delivered-To: devl at freenetproject.org
Received: from k-vort2.zd.guj.de (fwhide.guj.de [193.7.250.35])
        by hawk.freenetproject.org (Postfix) with ESMTP id E999158184
        for <devl at freenetproject.org>; Thu, 19 Apr 2001 09:16:07 -0700 (PDT)
Received: from hermes.guj.de (hermes.zd.guj.de [193.7.241.37])
        by k-vort2.zd.guj.de (8.11.3/guj-zz-1.4) with ESMTP id f3JGAUV03594
        for <devl at freenetproject.org>; Thu, 19 Apr 2001 16:10:30 GMT
Received: from drjava.de ([192.168.176.135])
        by hermes.guj.de (8.9.1b+Sun/8.9.1) with ESMTP id SAA10521
        for <devl at freenetproject.org>; Thu, 19 Apr 2001 18:10:29 +0200 (MET 
DST)
Message-ID: <3ADF0FC1.CAB29EAD at drjava.de>
From: Stefan Reich <[email protected]>
X-Mailer: Mozilla 4.73 [en] (WinNT; U)
X-Accept-Language: en
MIME-Version: 1.0
To: devl at freenetproject.org
Subject: Re: [freenet-devl] Crisis - Freenet Reliability and Performance
References: <002a01c0c8bc$af7f0d40$eb7436d2 at private> <20010419165331.B1209 
at hobbex.localdomain> <004701c0c8e7$ab889730$cbe636d2 at private>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: devl-admin at freenetproject.org
Errors-To: devl-admin at freenetproject.org
X-BeenThere: devl at freenetproject.org
X-Mailman-Version: 2.0.3
Precedence: bulk
Reply-To: devl at freenetproject.org
List-Help: <mailto:devl-request at freenetproject.org?subject=help>
List-Post: <mailto:devl at freenetproject.org>
List-Subscribe: <http://lists.freenetproject.org/mailman/listinfo/devl>,
        <mailto:devl-request at freenetproject.org?subject=subscribe>
List-Id: Discussion of information related to Freenet development 
<devl.freenetproject.org>
List-Unsubscribe: <http://lists.freenetproject.org/mailman/listinfo/devl>,
        <mailto:devl-request at freenetproject.org?subject=unsubscribe>
List-Archive: <http://lists.freenetproject.org/pipermail/devl/>
Date: Thu, 19 Apr 2001 18:18:09 +0200
X-Mozilla-Status: 8011
X-Mozilla-Status2: 00000000
X-UIDL: 3adbdd6c000000f5
Status: RO
Content-Length: 1721
Lines: 40

What I'm wondering in this context is... how can we be sure that Freenet isn't
fragmented? A few months ago, Gnutella suffered from fragmentation so badly that
it was virtually unusable. The main problem was de-facto fragmentation caused by
nodes with way too little bandwidth.

For Gnutella, it only got better since 2nd generation clients kill connections
to slow hosts (and immediately connect to other hosts). But still, I would
assume Gnutella consists of at least a handful fragments at any given point in
time.

With Freenet the situation is worse because its goals are more ambitious
(basically, any file should be requestable from anywhere). Since you're always
inserting into just one fragment, files won't be visible on other fragments
unless

-the files are requested at many different places AND

Reply via email to