On Jun 29, 2010, at 7:15 PM, Claudio Jeker wrote: > On Tue, Jun 29, 2010 at 06:59:28PM -0700, Mitchell Erblich wrote: >> >> On Jun 29, 2010, at 6:22 PM, Greg Mirsky wrote: >> >>> Dear Glen, >>> the OSPF uses UDP which is unreliable IP transport, not like TCP which is >>> used by BGP. You've noticed that some LLS (DD and LSA) have reliability >>> support on protocol level while others (Hello) don't. If Hello goes through >>> (at least one of three to keep adjacency up) then all other exchanges will >>> work, even with retransmissions. >>> >>> Regards, >>> Greg >>> >> >> Greg, >> >> Sorry UDP is NOT a IP transport. >> These last two words (IP transport) don't make sense. >> UDP is a protocol, that in many current environments >> is layered below the IP protocol. >> >> Yes, OSPF uses the UDP protocol. > > No it does not. OSPF uses its own proto 89 on top of IP. > OSPF has its own reliability layer to make sure that the LSDB stay in > sync. > > -- > :wq Claudio > _______________________________________________ > OSPF mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/ospf
Sorry, I misspoke. :( Some implementations of OSPF, can use UDP when certain OSPF capabilities are enabled and/or used. It is not a "normal" behaviour specified in the OSPF protocol and it is normally not used as a transport for OSPF. In one instance, UDP can be used when IP address/name and/or DNS names are resolved within the CLI. Or when some OSPF functions are executed when xmiting or recving OSPF protocol pkts. Mitchell Erblich _______________________________________________ OSPF mailing list [email protected] https://www.ietf.org/mailman/listinfo/ospf
