Re: Autonomous System Sanity Protocol

Pedro Marques <> Sat, 26 April 1997 17:57 UTC

Received: from by id aa24059; 26 Apr 97 13:57 EDT
Received: from cnri by id aa23947; 26 Apr 97 13:53 EDT
Received: from by CNRI.Reston.VA.US id aa11932; 26 Apr 97 13:53 EDT
Received: (roque@localhost) by (8.6.12/8.6.5) id KAA21019; Sat, 26 Apr 1997 10:50:21 -0700
Date: Sat, 26 Apr 1997 10:50:21 -0700
Message-Id: <>
From: Pedro Marques <>
To: Noel Chiappa <>
Cc:, ietf@CNRI.Reston.VA.US
Subject: Re: Autonomous System Sanity Protocol
In-Reply-To: <>
References: <>
Mime-Version: 1.0 (generated by tm-edit 7.105)
Content-Type: text/plain; charset=US-ASCII
Source-Info: From (or Sender) name not authenticated.

>>>>> "Noel" == Noel Chiappa <> writes:

    Noel>     From: Tony Li <>
    >> We need to move to a routing architecture where maps are
    >> distributed, *not* routing tables.

    Noel>     Exactly how does this prevent the exchange of bad
    Noel> information?

    Noel> Well, a full-scale explanation is a major tome (we can
    Noel> explore that on Big-I in more detail if you want), but
    Noel> *briefly*, the idea is that you can i) prevent lots of kinds
    Noel> of bad information, and ii) deal much better with the kinds
    Noel> you can't stop.

    Noel> For instance, use of public key cryptography can prevent
    Noel> anyone else from originating bad information about
    Noel> connectivity inside or to X - their map updates will not be
    Noel> correctly signed with X's private key. Only "auhorized"
    Noel> agents of topological entity X (i.e. those allowed to
    Noel> distribute maps or abstractions of X, outside X) have the
    Noel> key to sign map data about X.

s/map/BGP route/g
... and everything you said still holds.