The MPLS WG Archive

["Santosh Gupta" <santoshg@daewoo.dti.daewoo.co.kr>]

Hello

    I have some doubts regarding Path Vector TLV in LDP. Could someone help me.

1. Why do we need to have Path Vector TLV in Label Map Message? If we follow Downstream On demand Ordered Control Mode of Label Distribution then what could be the need for Path Vector TLV in Label Map Message. The same path was followed in Request Message.
2. The only use that I can think of Path Vector TLV in Label Map Message is when Egress LER didnt receive any Path Vector TLV/ Hop Count TLV so it sends back Label Map message so as to make sure that no loop is formed.
3. There are 2 ways in which an LSR could know it's HopCount. Either it receives a Label Request Message which carries it's previous Hop's "HopCount" or if it is 0 then the Label Map message ( which comes from DownStream peer ) will contain the "HopCount" once this LSP is formed till the Egress point. So is the LSR supposed to maintain 2 "HopCount" values for different directions?
4. How can each LSR know the no. of Hops in the LSP? Is it to be made from no. of entries in PathVector TLV when in Label Map message or their is some other means?

thanx in advance.

 

santosh