[26109] in North American Network Operators' Group
RE: Verio Decides what parts of the internet to drop
daemon@ATHENA.MIT.EDU (Alex P. Rudnev)
Sun Dec 5 03:10:03 1999
Date: Sun, 5 Dec 1999 10:48:05 +0300 (MSK)
From: "Alex P. Rudnev" <alex@virgin.relcom.eu.net>
To: smd@clock.org
Cc: rmeyer@mhsc.com, nanog@merit.edu, randy@psg.com, tony1@home.net
In-Reply-To: <19991204232227Z27421-10571+8@cesium.clock.org>
Message-ID: <Pine.SUN.4.10.9912051043031.29042-100000@virgin.relcom.eu.net>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Errors-To: owner-nanog-outgoing@merit.edu
I just agree. It was only 4x improvement.
I think Tony did the excellent answer. I want to highlight the danger of such,
non-coordinated extra filtering done by one ISP. It's really dangerous for the
Internet. On the other hand, no one can allow /32 or /28 prefixes, and no one
like (but everyone allow) /24 and /23 prefixes. And I knew a few cases when such
prefixes was injected into the global network in purpose (it was the only chance
to get the riught result quickly).
In addition, such filtering cause extra usage of the adress space, and I prefere
to see less limits in the prefix sizes because the exhausting of the address
space is real danger (not as the CPU or router memory lack).
On Sat, 4 Dec 1999 smd@clock.org wrote:
> Date: Sat, 4 Dec 1999 15:22:19 -0800
> From: smd@clock.org
> To: alex@virgin.relcom.eu.net, rmeyer@mhsc.com
> Cc: nanog@merit.edu, randy@psg.com, tony1@home.net
> Subject: RE: Verio Decides what parts of the internet to drop
>
>
> | The memory for the routing tables was a deal just about 2 years ago; this
> | became easier to maintain big tables today (when routers can be easily upgraded
> | to 256 MB RAM).
>
> So, Alex, this is a 4x improvement over the situation in 1995,
> when 64Mb was possible but rare...
>
> Sean.
>
Aleksei Roudnev,
(+1 415) 585-3489 /San Francisco CA/
