Re: [Lsr] LSR WG Adoption Poll for "Flexible Algorithms: Bandwidth, Delay, Metrics and Constraints" - draft-hegde-lsr-flex-algo-bw-con-02

[Aijun Wang <[e-mail]>]

Hi, Tony:

7.      The document introduces a new Generic Metric type called Bandwidth metric. I’ve been trying to follow some of the discussion related to this on the mailing list – about it being cumulative or not. I am perhaps somewhat confused by those discussions. The OSPF/ISIS SPT computation has always worked with cumulative link (and prefix) metrics. If the computation for the Generic Metric of this new type b/w is not going to be cumulative (I thought it is – but not very clear anymore), then the document needs to describe the computation algorithm. Is it then hop count based? Perhaps I am missing something very basic here and if so, please point me to the text in the draft.

 

 

I’m sorry if this has been confusing. My understanding is that the metric is cumulative. Others had other expectations.

 

When there are multiple links with the same bandwidth, and thus the same metric, then the total path metric becomes (link metric) * (number of links).

[WAJ] Such cumulative may be problematic in some situations and would need the manual intervention to “override” the automatic calculation “total path metric” to achieve the desired results. (please refer to the example in https://mailarchive.ietf.org/arch/msg/lsr/wWoGgwf-Nch0_VxjczZBpLFXyos/

The difficulty is to how to find the unexpected paths in complex topology.

 

My suggestion is still not introduce such non-cumulative metric to cumulative based SPF calculation process.

 

 

Best Regards

 

Aijun Wang

China Telecom

 

 

 

 

Regards,

Tony