On Wed, 21 Sep 2005, Jari Arkko wrote:
Presumably all you'd need to do is to look at all packets that have protocol = icmpv6 (despite whether they are addressed to you or not). You might filter further based on the type of message, but I think we'd already be in the neighborhood of feasible implementation. And the actual intelligence might be either in a central node or in the switches themselves.
Uhh, on a switch which is supposed to be forward, line-rate, dozens of gigabytes of traffic per second this is really NOT all that simple. But the problem already exists to some degree..
But nonetheless, the draft is not sufficiently convincing on why exactly we should bother with non-backwards-compatible changes. Such changes are really major, and would give a strong indication that the problem should be changed in some other way. If you intend to continue seeking a solution to this problem, I'd put a lot more emphasis on the definition of the problem and evaluation of the current tools (and why those are not sufficient).
-- Pekka Savola "You each name yourselves king, yet the Netcore Oy kingdom bleeds." Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
