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

Re: [RPRWG] Class A and B Guarantees



 
Necdet,
 
 
 
NU: Did you run any simulations showing the priority inversion happening while adding classA1 (when stqFullThreshold - stqHighThreshold > RTT * rateA1? 
 
Not yet, but A0, A1 or class B traffic should not make a difference in
the scenario, they all have the same guaranteed bandwidth problem.
I don't understand the last part of your question, but the RTT actually
does not play a role in the example scenario. The problem is the amount
of queued traffic on the transit path since this cannot be stopped by FCMs.
So it doesn't matter if these FCM's come quickly or not.
 
NU: Did you implement shaperD and reserved classA0 bandwidth all around the ring?

Yes, I can dynamically make reservations for class A0, A1 and B traffic. I'm not sure what you mean by "all around the ring".

Simulations are great, but this scenario is not so difficult to comprehend:
If 100 stations start sending class C packets to one station, and the STQs are all empty, then
all these station may add local traffic until their stqLowThreshold is reached. The amount of
traffic in all these STQs  is tremendous and will produce a very long traffic stream at 100% of
the line rate. And the worst thing: it cannot be stopped by FCMs!
Stations receiving this stream, which also have guaranteed bandwidth flows, can get in trouble,
because their STQ may overload.

Jon

 

 


 

 
 
 
----- Original Message -----
Sent: Tuesday, April 22, 2003 8:24 PM
Subject: Re: [RPRWG] Class A and B Guarantees

Jon,

Please see my comments in line.

Thanks.

Necdet

Jon Schuringa wrote:

Dear all, 
I posted  a comment (#33) at the Dallas meeting about bandwidth
guarantees: In my opinion, bandwidth agreements cannot always
be guaranteed. The comment was rejected because it was addressed to the wrongclause. Although at the wrong address, I got the answer that thestatement in my comment is incorrect, but without any explanation. Since then I had discussions with several people, and checked my
simulations with another simulation tool (ns2). As before, I strongly
believe this to be a serious technical concern, and therefore post it
here to the mailing list.  
The problem in short:
STQ's can reach the stqFullThreshold in scenarios where both class C
and class A traffic flows. As a result, the STQ gets precedence over
all locally sourced traffic, so that class A (and B) traffic has to wait,
causing bandwidth and jitter problems.
The STQ can get that full because fairness messages cannot stop
packets that already have been transmitted by other stations, but did
not yet arrive at the local station. This amount of packets that is on the
transit path can be very large since it is the sum of all packets in the
STQs on the transit path. This is also the reason why larger STQs
do not solve the problem. So basically what happens in the problem scenarios is that:
1)  the local station (S) receives class C packets at 100% of the line rate.
     All these packets need to be forwarded by station S
2)  Station S transmits guaranteed class A (local) traffic at some rate x,
     so the local STQ grows (at rate x).
3)  Station S advertises a fair rate unequal to FULL_RATE once the STQ
     exceeds the stqLowThreshold
4)  All other stations see the advertized rate and limit their "add" traffic.
    This however does not directly prevent that station S gets less than
    100% line rate, because there is still transit traffic that needs to be
    forwarded by all stations. These stations empty their STQs.
5) If the class A rate x and the number of STQs are "large enough", the
    STQ in station S will reach its stqFullThreshold and priority inversion    is the result. Note that the potential problem scenarios are realistic hub-scenarios, not
"pathological cases".

NU: Did you run any simulations showing the priority inversion happening while adding classA1 (when stqFullThreshold - stqHighThreshold > RTT * rateA1?
 
  A detailed description and an example scenario can be found here:
 http://grouper.ieee.org/groups/802/17/member/draftballots/d2_1/refs/js_issues_1.pdf
This document contains other issues as well.
  Opinions?
 

NU: Did you implement shaperD and reserved classA0 bandwidth all around the ring?
 
 
  Best regards,Jon -----------Jon SchuringaInstitute of Communication Networks
Vienna University of Technology
Favoritenstraße 9/388
A-1040 Vienna

+43/1/58801-38814
www.ikn.tuwien.ac.at