RE: [RPRWG] Scheduling

[Leon Bruckman <leonb@xxxxxxxxxxxxx>]

Title: RE: [RPRWG] Scheduling

I agree with William, 3 CoS supported by the MAC layer provide a good compromise between the types of services and the implementation. Further classification can be performed by higher layers, and should be left out of the scope of the MAC definition.

 

Leon Bruckman

-----Original Message-----
From: William Dai [mailto:wdai@xxxxxxxxxxxx]
Sent: Thursday, April 05, 2001 2:32 AM
To: stds-802-17@xxxxxxxx
Subject: Re: [RPRWG] Scheduling

Again, I'm talking about behavior at the shared media MAC layer.

To be politically correct, the upper layer I mentioned DOES NOT 

necessarily mean L3.

 

There are subtle differences between shared media MAC behavior 

and Hop-by-Hop behavior.

 

William Dai

----- Original Message -----

From: Sanjay Agrawal

To: 'William Dai' ; stds-802-17@xxxxxxxx

Sent: Wednesday, April 04, 2001 5:01 PM

Subject: RE: [RPRWG] Scheduling

How about overcommitted class.

This makes it 4. Should we have provision for other classes.

If Diffserv 3 bits of COS (next 3 bits are for drop presidence ), MPLS COS (3 bits), and 802.1Q (3 bits)are

in conformance, why should RPR be any different.

 

We should even leverage from the code points which Diffserv has defined.

 

101 for EF

100 for AF4

011 for AF3

010 for AF2

001 for AF1

000 Best Effort

 

 

 

-Sanjay K. Agrawal