The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2004-Mar> msg00012



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

Routing plane covergence of Distance Vector Protocols.

  • From: Spice Sylvia <falsesylvia@yahoo.co.uk>
  • Date: Sun, 7 Mar 2004 18:49:28 +0000 (GMT)

Given a closed graph (non tree),

update/withdrawl-->R1----R2---R3----R4---->update/withdrawal
                               |-- R5-----R6---|


(hope the ASCII bit comes fine, else it simply is R1
connects to R5 and R5 to R6 and then R6 to R4),

If one can represnt the routing algorithm's decision
process as a function F(input_at_time_t) at each node,
and
a. has the ability to modify the input function(may
not be impulse)
b. choose the function F (which is the same at each
node)

can one do away with principals like Count to Infinity
in RIP(Bellman-Ford) and get a way that the topology will converge
to stability when any input is applied at R1?
Purely an academic question.
 
-Sylvia


Yahoo! Messenger - Communicate instantly..."Ping" your friends today! Download Messenger Now