The MPLS WG Archive

Cell Relay Retreat>MPLS WG Archive>month:2000-Oct> msg00556



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

[IP-Optical] Joint Routing

  • From: Mark Stewart <Mstewart@nexen.com>
  • Date: Fri, 27 Oct 2000 11:42:24 -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>

Hi Dave

Well I certainly hope no one is proposing to mandate a centralized mechanism to implement joint routing. It's quite possible to do a source based solution which simply invokes dykstra twice instead of once and then has the same scaling and synchronization issues as OSPF.

ciao

mark

David Allan wrote:

 

True, they are separate concepts. Centralzied computing is A solution to the problem of ensuring a synchronized and authoriative view of the network which is what I would assume joint routing requires. So what I am really discussing is the scaling limitations that a synched and comprehensive network view has regardless of how the algorithm is implemented, especially as the dynamic behavior of the network increases with time.

Dave

    -----Original Message-----
    From:   Mark Stewart [SMTP:Mstewart@nexen.com]
    Sent:   Friday, October 27, 2000 11:02 AM
    To:     Allan, David [CAR:NS00:EXCH]
    Cc:     alchiu; 'Yakov Rekhter'; ip-optical; mpls; sc; xuyg; yxue
    Subject:        Re: [IP-Optical] Joint Routing was RE: Optical link bundling. Was Re:  DraftMinutes From Pittsburgh

    Hi Dave

    Why are we equating joint routing with centralized computation? Few (if
    any) members of the IETF would dispute that centralized computation has
    nasty scaling problems; but I thought we were discussing joint routing.

    Mark

    <snip>