Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email firstname.lastname@example.org
From: Alex Zinin (azininNEXSI.COM)
Date: Tue Sep 04 2001 - 20:29:01 CDT
I understood your point.
Still, the flooding and LSDB maintenance algorithms
are not (and I think should not be) aware about the specifics
of certain LSA contents, they are generic for both non-TE
and TE information. An LSA with a missing mandatory
sub-TLV should be accepted (i.e., no changes should be
introduced into the flooding algo), installed in the LSDB, and
then the client (the TE component) should decide whether
the contents is valid for TE-specific processing (CSPF)
-- Alex Zinin
Tuesday, September 04, 2001, 6:00:32 PM, Singh, Gurpreet wrote:
> Thanks for the info.
> I mean if some mandatory Sub-TLV is not present then the router could > consider it as an invalid LSA and drop it. I am not sure if the metric > sub-tlv is made a mandatory TLV then the Router should still consider it as > a valid LSA when it receives one with no metric-LSA or not. Some signalling > protocols do have this kind of behaviour of droppping the Packet when some > mandatory object is not present.
> -----Original Message----- > From: Alex Zinin [mailto:azininnexsi.com] > Sent: Tuesday, September 04, 2001 8:48 PM > To: Singh, Gurpreet > Cc: OSPFDISCUSS.MICROSOFT.COM > Subject: Re: Question on draft-katz-yeung-ospf-traffic-05
> You're not supposed to drop an LSA because you believe > some sub-TLV is not there. The LSA should be accepted and > installed in the LSDB. Whether you're going to use it in > CSPF and how is a different issue.
> -- > Alex Zinin
> Tuesday, September 04, 2001, 8:43:12 AM, Singh, Gurpreet wrote:
>> Since the draft only says that Link Type and the Link ID sub-TLVs are > the >> only mandatory ones, what should be the behaviour of the Router when it >> receives a TE Opaque LSA with only Link ID and Link Type Sub-TLVs in the >> Link TLV. Shoiuld it drop the packet or accept it as a valid packet.
>> -----Original Message----- >> From: Kireeti Kompella [mailto:kireetiJUNIPER.NET] >> Sent: Tuesday, September 04, 2001 11:30 AM >> To: OSPFDISCUSS.MICROSOFT.COM >> Subject: Re: Question on draft-katz-yeung-ospf-traffic-05
>>> 1) In case the traffic engineering metric sub-tlv is not present in the >> link >>> tlv, does the traffic engineering SPF calculation have to rely on the >>> standard OSPF link metric ?
>> There is no requirement that there be a "standard OSPF link" advertised >> along with the opaque TE TLV. Thus this is not an option.
>> My inclination is to make the TE metric a mandatory sub-TLV of the >> TE Link TLV.