Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
From: Manav Bhatia (manavSAMSUNG.CO.KR)
Date: Tue Sep 11 2001 - 23:56:06 CDT

  • Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

    Hi Alex,
    I'm really not sure which problem are you reffering to.
    Say a router crashed and is rebooting while the network is
    recalculating paths to it. When it starts its link state protocol
    back up, it must somehow indicate that it needs to reinitialize
    its sequence number to the last number it gave all of the other
    routers to allow for coherence. Therefore, it can announce
    paths with a sequence number in a special "initialization set". This
    initialization set will tell the other routers that this router needs the
    sequence where it left off. This is the "lollipop sequence" idiom.
    In this the normal sequence numbers keep churning around the
    finite sequence space while reinitialization takes place in a short
    linear sequence space.

    So, what advantages do we get in using the linear number


    ----- Original Message -----
    From: <azininNEXSI.COM>
    Sent: Wednesday, September 12, 2001 4:57 AM
    Subject: Re: about the sequence number

    > zhang
    > It was lollipop in OSPF v1, but this approach is known
    > to have a problem (known as S1, S2, S3). The linear
    > sequence space does not have this problem.
    > I think John explains it in his book.
    > --
    > Alex Zinin
    > Tuesday, September 11, 2001, 2:44:45 AM, ? ?? wrote:
    > > Hi all,
    > > In RFC2328, the sequence number of LSA is linearly
    > > ordered(Page 120),i.e., it is linearly number space,
    > > why it is not lollipop-shaped sequence number space?
    > > Thanks!
    > > Regards,
    > > zhang