Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [8023-CMSG] Problem statement

Title: Problem statement
Hi Brad/David,
 
    1. Yes, by priorities, I was referring to .1p. Classifying flows
       acc/g to priorities and scheduling queues with algo (strict,
       WRR..)
    2. However, keeping in mind that since 802.3 does not specify
       "differentiation" or "priorities" for flow, NICs need to have
       mechanism mentioned in 802.1 (for bridges). Probably part of
       OS stack. (I am not sure of standardization here).
    3. If "Hi Pri" traffic in itself does not cause congestion, then
       it can be carried through the network without any noticeable
       impact of congestion if appropriate classification/scheduling
       is performed at all the nodes on the path. I have yet to get
       solid data about "bunching up" effect on these flows.
    4. Oversubscription can not be solved by prioritizing. However
       even in oversubscribed case "hi pri" traffic (and other classes)
       can get through congested network: depends upon amount of BW
       it needs. (E.g. If switch is having 50G switching BW, and 60G
       total traffic is pumped into it, and Hi Pri traffic is only 5G,
       then with appropriate scheduling, 10G "Lo Pri" traffic will only
       be dropped/delayed. Leaving "Hi Pri" traffic practically safe).
        -> Again, I would like to see data about "bunching up" effect
       that has been discussed in some forums.
    5. About what CMSG is trying to accomplish: Brad,you have summed it
       up pretty well in another email.
 
Thanks,
- Manoj

From: owner-stds-802-3-cm@listserv.ieee.org [mailto:owner-stds-802-3-cm@listserv.ieee.org] On Behalf Of Booth, Bradley
Sent: Tuesday, September 14, 2004 11:04 AM
To: STDS-802-3-CM@listserv.ieee.org
Subject: Re: [8023-CMSG] Problem statement

Manoj,
 
Can you clarify "differentiation" or "priorities"?  I believe you mean those features within 802.1, not being added to 802.3.  Is that correct?
 
Thanks,
Brad

From: owner-stds-802-3-cm@listserv.ieee.org [mailto:owner-stds-802-3-cm@listserv.ieee.org] On Behalf Of Wadekar, Manoj K
Sent: Monday, September 13, 2004 4:56 PM
To: STDS-802-3-CM@listserv.ieee.org
Subject: Re: [8023-CMSG] Problem statement

No, CMSG focused primarily on "oversubscription" issue.
["Transient" being addressed by "differentiation" or "priorities"].
 
 
Thanks,
- Manoj

From: owner-stds-802-3-cm@listserv.ieee.org [mailto:owner-stds-802-3-cm@listserv.ieee.org] On Behalf Of David V James
Sent: Monday, September 13, 2004 2:41 PM
To: STDS-802-3-CM@listserv.ieee.org
Subject: Re: [8023-CMSG] Problem statement

Hmm,...
 
1) I had thought the primary reason for congestion management was to avoid
the short-term problem of loss of traffic during coincidental peaks in traffic.
 
The term "oversubscription" seems to imply a long-term flow control solution.
I suppose that's OK if the original intent of (1) was misperceived or has changed.
 
DVJ
 

David V. James
3180 South Ct
Palo Alto, CA 94306
Home: +1.650.494.0926
      +1.650.856.9801
Cell: +1.650.954.6906
Fax:  +1.360.242.5508
Base: dvj@alum.mit.edu 

-----Original Message-----
From: owner-stds-802-3-cm@listserv.ieee.org [mailto:owner-stds-802-3-cm@listserv.ieee.org]On Behalf Of Booth, Bradley
Sent: Monday, September 13, 2004 1:59 PM
To: STDS-802-3-CM@listserv.ieee.org
Subject: [8023-CMSG] Problem statement

Greetings,

I wasn't able to attend the CMSG meeting in July, due to being a little busy in 802.3an, but I was looking at the problem statement that I believe was adopted by the SG.  I was a little concerned that the statement only mentioned 802.3 MAC Clients and nothing about the 802.3 MAC itself.  I was wondering if the following problem statement would still be palatable to everyone:

"802.3 MAC Clients need the ability to communicate, via 802.3 MACs, congestion information to avoid oversubscription."

Thoughts?  Feedback?

Thanks,
Brad