[34173] in North American Network Operators' Group
Re: BGP and anycast
daemon@ATHENA.MIT.EDU (hardie@equinix.com)
Tue Jan 30 13:22:32 2001
Message-Id: <200101301819.KAA23869@nemo.corp.equinix.com>
To: marcs@znep.com (Marc Slemko)
Date: Tue, 30 Jan 2001 10:19:40 -0800 (PST)
Cc: hardie@equinix.com, ymanon@yahoo.com (Swede),
MSchoenecker@yipes.com (Mike Schoenecker), nanog@merit.edu
In-Reply-To: <Pine.BSF.4.20.0101300949400.84647-100000@alive.znep.com> from "Marc Slemko" at Jan 30, 2001 09:58:21 AM
From: hardie@equinix.com
Reply-To: hardie@equinix.com
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Errors-To: owner-nanog-outgoing@merit.edu
A good point. This document assumes a DNS context, and thus that the
UDP request and response are self-contained. I will attempt to make
it more clear in the text, but this is exactly the sort of caution I
was trying to get at: do not assume that a hack that works in some
circumstances for the DNS will work for other services.
regards,
Ted Hardie
>
> On Tue, 30 Jan 2001 hardie@equinix.com wrote:
>
> > One potential problem with using shared unicast addresses is that
> > routers forwarding traffic to them may have more than one available
> > route, and those routes may, in fact, reach different instances of
> > the shared unicast address. Because UDP is self-contained, UDP
> > traffic from a single source reaching different instances presents
> > no problem. TCP traffic, in contrast, may fail or present
>
> That should be a little more precise.
>
> TCP packets can not (for all practical purposes when dealing with "normal"
> clients) be self contained.
>
> UDP packets are self contained, from the network view.
>
> But that does not mean that a particular protocol implemented on top of
> UDP will necessarily still be self contained, merely that it is possible
> for it to be.
>
