[1260] in linux-net channel archive
Re: ip aliasing support ...
daemon@ATHENA.MIT.EDU (Matthias Urlichs)
Fri Oct 27 11:21:34 1995
To: submit-linux-dev-net@ratatosk.yggdrasil.com
From: urlichs@smurf.noris.de (Matthias Urlichs)
Date: 25 Oct 1995 18:19:10 +0100
In linux.dev.net, article <JAA00819@sys3.pe1chl.ampr.org>,
linux-vger@wab-tis.rabobank.nl writes:
>
> (and note the code not even covers non-contiguous netmasks...)
>
There are no noncontiguous netmasks. Their status is somewhere between
"strongly discouraged" and "outlawed" in current RFCs.
> Having binary-searched arrays instead of the hashed lists could be
> faster in lookups in some cases (I am not convinced), but certainly
> is less efficient when updating the table.
>
I too think a hash is a better idea. (32 hashes?) Or you could optimize by
grouping, say, four bits. That'd mean eight hash lists (of varying sizes,
of course), but a specific route would have to be entered up to eight
times. I think that'd be a good balance between access speed and memory
overhead, and most routes are likely to fall on a four-bit boundary anyway
because the old class-A/B/C addresses do.
> To keep the process of route lookup efficient, NOS uses a route cache.
> (originally one entry, I have extended it to have as many entries
> as the hashed lists).
Ditto.
--
I always wanted to marry someone who was tall, handsome, and rich.
Three out of three I gave up on.
-- Woody Allen, "Radio Days"
--
Matthias Urlichs \ XLink-POP Nürnberg | EMail: urlichs@smurf.noris.de
Schleiermacherstraße 12 \ Unix+Linux+Mac | Phone: ...please use email.
90491 Nürnberg (Germany) \ Consulting+Networking+Programming+etc'ing 42
PGP: 1B 89 E2 1C 43 EA 80 44 15 D2 29 CF C6 C7 E0 DE
Click <A HREF="http://smurf.noris.de/~urlichs/finger">here</A>.
