Steven M. Bellovin wrote: >In message <[EMAIL PROTECTED]>, Hiroki Ishibashi writes: > >>Now, OSPFv3. Since it is much complicated than any RIPng, of course, >>and it has capability to run multiple processes by nature, we decided to >>run an OSPFv3 process per site. Still, we could have handle site routing >>with single OSPFv3 process. Complication is the most dominant factor in our >>decision. For example, what if someone is sending an Intra-Area-Prefix-LSA >>which contains both global and site-local prefixes. Receiving OSPFv3 router >>might need to decompose LSA into two LSAs which contain global and site-local >>separately. BUT, this is not allowed. I believe that we need to make >>agreements on this point. >>One of our concerns on multi-process implementation is that >>global prefixes need to be redistributed among OSPFv3 process once >>sites are defined. >> > >I'm still confused. If a packet arrives on a site-enabled interface, >addressed to multicast address AllSPFRouters, and with protocol number >89 (OSPF), to which process is it delivered? Does something actually >peek inside the packet to see if it's advertising global or site-local >addresses when making the dispatching decision?
"2.4. Explicit support for multiple instances per link" in RFC2740 is the one to be used to identify a process (instance) to which packet are delivered. Not necessary to peek into LSAs for the dispatching decision. Hiroki Ishibashi [EMAIL PROTECTED] -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
