The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2000-Oct> msg00570



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

[IP-Optical] Joint Routing

  • From: "Frank Hujber" <fhujber@hotmail.com>
  • Date: Fri, 27 Oct 2000 14:11:15 -0400
  • Cc: "alchiu" <alchiu@research.att.com>, "'Yakov Rekhter'" <yakov@cisco.com>, "ip-optical" <ip-optical@lists.bell-labs.com>, "mpls" <mpls@UU.NET>, "sc" <sc@tellium.com>, "xuyg" <xuyg@lucent.com>, "yxue" <yxue@UU.NET>
  • X-OriginalArrivalTime: 27 Oct 2000 18:09:54.0246 (UTC) FILETIME=[1B5C4E60:01C04041]
  • X-Originating-IP: [12.124.181.202]

ÿþ<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META content="text/html; charset=unicode" http-equiv=Content-Type> <META content="MSHTML 5.00.2314.1000" name=GENERATOR> <STYLE></STYLE> </HEAD> <BODY bgColor=#ffffff> <DIV><FONT face=Arial size=2>Dave, Mark, et al,</FONT></DIV> <DIV>&nbsp;</DIV> <DIV><FONT face=Arial size=2>I've been watching this one for a while and I cannot let it go by without offering my viewpoint. Which is...</FONT></DIV> <DIV>&nbsp;</DIV> <DIV><FONT face=Arial size=2>There ought to be a compromise between the distributed approach and the centralized approach whereby the ingress (or maybe the egress) router has a picture of the topology and from that makes the explicit route. A distributed approach will not assure diverse paths and will force the computation to be delayed until the fault has occurred and some fault analysis algorithm takes some time to specify which link or node to avoid. A fully centralized approach clearly has scaling problems. I agree, keeping the topology picture uniform will be a challenge. If it wasn't it would've been done already. However, consider that the amount of time that a lightpath is likely to exist may be longer than the average datagram. The time constant associated with router updates may be longer as well, so this eases the burden a bit when it comes to updating the topology picture (i.e. the&nbsp;optical link state.)</FONT></DIV> <DIV>&nbsp;</DIV> <DIV><FONT face=Arial size=2>I've seen hints of this compromise passing by over the past few days, but somehow I think it keeps getting lost in the definitions of "centralized" and "distributed." It just doesn't seem to me like a three-sigma answer is the right one.</FONT></DIV> <DIV>&nbsp;</DIV> <DIV><FONT face=Arial size=2>Frank Hujber</FONT></DIV> <DIV><FONT face=Arial size=2><A href="mailto:fhujber@hotmail.com">fhujber@hotmail.com</A></FONT></DIV></BODY></HTML>