Hi Adrian, Historically, the draft/RFC is titled router capabilities but you'll note that the actual name of the LSA is the RI (Router Information LSA). I do believe that the system name fits in the category of information and it WILL be used in OSPF for routing - not directly for calculating OSPF routes but for debugging networks when routing doesn't work. Hence, IMHO, it satisfies both the criteria for inclusion in the default instance and the criteria for inclusion in the OSPF RI LSA. Like most of the other information in the RI LSA, I'd expect it be fairly static.
Thanks, Acee On Mar 20, 2008, at 5:19 PM, Adrian Farrel wrote: > Acee, > > I missed the last two OSPF meetings because of agenda clashes, and > I haven't > seen much discussion on the list. > > I support the idea of being able to exchange a textual name for > routers, and > having this I-D in the working group may provide a way to get the > mechanism > right. > > I am worried as Nitin was about the idea of using the RI LSA since > the name > of a router is not a router capability. And we wouldn't want the > 255 byte > reflooded every time some other routing capability was modified. > Indeed, the > consumer of the router name is unlikely to be the IP routing and > forwarding > component. So the information should be moved to some other LSA > that can be > handled with different refresh rules and potentially moved into a > separate > OSPF instance per > > I note that you said >> This isn't the type of information we are envisioning needing >> to be split out. It is only a single name and will be an optional >> TLV for an existing LSA. The separate instance would be >> more for flooding larger volumes of non-routing information >> associated with new applications (once we have this instance, >> it also would be nice to move some existing opaque LSAs to >> it as well). > > The fact that it is optional for an LSA, doesn't seem relevant. > Once it is > in the LSA it has to be flooded. > > Additionally, this could well be the thin end of the wedge. > Wouldn't it be > useful for network management tools if the routers flooded some > geographical > coordinate? Wouldn't it be handy for operators if the postal > address and > phone number of the maintenance site were flooded? Where will this > lead? > > Take the clean approach now. Recognise that information not used > for IP > routing does not belong in the primary instance of the IGP. > > So, I guess I give you a qualified "yes" to your question. Make > this a WG ID > if the debate about moving the information out of the RI LSA is > still open. > > Cheers, > Adrian > > ----- Original Message ----- > From: "Acee Lindem" <[EMAIL PROTECTED]> > To: "OSPF List" <[email protected]> > Sent: Thursday, March 20, 2008 8:02 PM > Subject: [OSPF] Dynamic Hostname Exchange Mechanism for > OSPF -draft-venkata-ospf-dynamic-hostname-02.txt > > >> This draft was presented in Vancouver and with an update in >> Philadelphia. At both WG meetings there was interest in making this a >> WG document and no opposition. Is there any one now who has a >> compelling reason why this should not be a WG document? If not, I >> plan to have the authors submit it as a WG document on Monday. Note >> that there will be plenty of time to comment on the technical >> contents of the draft. >> _______________________________________________ >> OSPF mailing list >> [email protected] >> https://www.ietf.org/mailman/listinfo/ospf >> > > > _______________________________________________ > OSPF mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/ospf _______________________________________________ OSPF mailing list [email protected] https://www.ietf.org/mailman/listinfo/ospf
