North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: RFC 1918

  • From: Greg A. Woods
  • Date: Fri Jul 14 20:14:11 2000

[ On Friday, July 14, 2000 at 12:01:25 (-0400), Shawn McMahon wrote: ]
> Subject: RFC 1918
>
> Some naughty person in either BBNPlanet or DeltaCom needs to re-read RFC 1918:

If only BBNplanet and DeltaCom (and OARnet, and [email protected], etc.,) were
the only offenders....

I've got megabytes full of log files showing guilty parties and that's
just from the very few nimrods who try to connect to my measly
almost-zero-content servers!  I've got even more megabytes of such crap
from "owned" M$ boxes that are trying to scan my network, and I happen
to know that they're not all just my @Home neighbours either.

Of course not all network operators are fully and directly guilty of
such abuses -- some are using broken equipment/software that leaks this
kind of crap....  Still, they should know better.

If I could only send a million-volt, mega-joule, packet back to every
firewall that uses an RFC1918 source address to try and tell me that I'm
not allowed to do IDENT queries to some server behind it that has
already connected to me....

Since I now have a couple year's experience with filtering all RFC-1918
addresses either at the borders or on servers in various situations I
can attest to the fact that one of the biggest problems with trying to
use RFC-1918 properly in an enterprise situation is that it's damn hard
to get everything to work correctly while at the same time honouring the
letter and the spirit of the restrictions in RFC1918.  I say this just
to provide yet one more datum to show why ISPs should *NEVER* use
RFC-1918 addresses in any of their public infrastructure, not ever.

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <[email protected]>      <robohack!woods>
Planix, Inc. <[email protected]>; Secrets of the Weird <[email protected]>