From: Charles Lindsey (chl@clerew.man.ac.uk)
Date: Thu May 06 2004 - 09:13:04 CDT
In <87zn8n8sse.fsf@windlord.stanford.edu> Russ Allbery <rra@stanford.edu> writes:
>Charles Lindsey <chl@clerew.man.ac.uk> writes:
>> But there might be problems due to the different Injection-Dates if the
>> multiple injections were separated in time
>Could you given an example of those problems?
Imagine a dumbell shapped Usenet, with well connected ends A and B,
and a slow channel in between that introduced delays in the propagation
between them (a little bit like it was 15 years ago, when the only
Usenet connection between Europe and the USA was a bit of spare capacity
on one NATO satellite).
An article is injected twice:
To a site in A:
Subject: foo
Date: 01 Jan 2003 10:00:00 ...
Injection-Date: 01 Jan 2003 10:01:00 ...
Message-Id: <12345@example.com>
.........
To a site in B:
Subject: foo
Date: 01 Jan 2003 10:00:00 ...
Injection-Date: 03 Jan 2003 10:01:00 ...
Message-Id: <12345@example.com>
.........
Observe the delayed Injection-Date in B (maybe it was carried across the
Atlantic on a laptop).
But even so, the article still had not propagated from the original A site
to B, so the 2nd version would rapidly propagate around the sites in B,
and versions which eventually arrived from A would be discarded as
duplicates by machines in B.
Finally, copies of the 2nd version try to propagate back across the
bottleneck towards A, and when they arrive there the articles have already
expired from the A machines, and so they accept the 2nd version because
its (late) Injection-Date seems OK.
A fairly improbable scenario, and not much that can be done about it if we
accept the Injection-Date proposal (which we should, because it should
improve things in just about every other scenario).
N.B. No reinjection occurred in any of the above.
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl Email: chl@clerew.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5