Re: [CCAMP] Comments to draft-wang-ccamp-latency-te-metric-01

[[e-mail]]

Hi Frederick,

Pls see inline.

Xihua Fu




Frederick Robinson <[e-mail]> 
2010-11-06 下午 10:04

收件人
[e-mail], [e-mail], [e-mail]
抄送

主题
Comments to draft-wang-ccamp-latency-te-metric-01






Dear Authors,

I agree with the requirement of communicating latency as a TE performance 
metric especially in OTN. I have some questions to the latency.
How to measure the latency of a single link? Perhaps you need to clarify 
it.
[Xihua] How to measrue the latency is out scope of this document. In OTN 
we may use the test signal or LMP protocol to measure the latency of link. 
In MPLS-TP we may use OAM packet to measure it. 
Do you really want to consider the latency of nodes for the path 
computation? It is very diffcult to measure it.
[Xihua] Yes, it do. We can configure the latency of node into control 
plane. Data plane may provide a mechnisom
I don't think you need to trigger the restoration/re-route because total 
latency of LSP has been affected by the linke latency. 
[Xihua] There is latnecy agreement between operator and its user (especial 
private line user). If the provisioned latency of pivate line could not 
meet the latency SLA again, operator should switchover the traffic to 
another LSP which meet the latency SLA. 

Regards,
Frederick