[ntp:questions] Detecting bufferbloat via ntp?

E-Mail Sent to this address will be added to the BlackLists Null at BlackList.Anitech-Systems.invalid
Wed Feb 9 23:58:45 UTC 2011


Dave Täht wrote:
> John Hasler <jhasler at newsguy.com> writes:
>> BTW there are queuing algorithms that work well with
>>  large buffers.
>
> Few of which appear to be in use.

If tools {all the many flavors of Bandwidth Management,
 Traffic Shaping, Congestion Avoidance, Queuing, ...
 {QOS & AQM (F|I|E|C|B|W|G|S|...)RED / (SF)Blue / REM / PI(P|D),
  (W|F|B)ECN, EFCI, ICMP SQ, ...} to deal with the
 issue are available, and some are not being used,
 that sounds like a problem those IT departments / NOCs
 can easily solve if they wanted to.


Why isn't AQM being leveraged more often, to make the best
 of the situation?

 I suspect a reason at many ISPs, (and IT departments,
   End Users), is they really don't care enough;

  Another reason might be that some of the AQM options
   just result in different packets being dropped.
    {I heard managers say, "Why throw away perfectly good
      packets when _our_ devices buffers aren't even full yet".}

  If they are meeting their contract requirements,
   why improve that issue, when they have innumerable
   other issues that affect their customers, that they
   also seem to care too little about.


I notice you run Ubuntu?  It appears to support ECN,
 {and RED (and other flavors) in ALTQ, CAR, ...)}.

  Do you have AQM/ECN enabled in your routers (core & border),
   and in your PC OS(s)?

    If not, why not?

  Then all you have to do is look for both of the ECN bits
   in the IP DiffServ Field (TOS byte) being on,
   and you know you have congestion.

    {Passive, & doesn't require NTP involvement.}


This is certainly drifting off topic for this group.
-- 
E-Mail Sent to this address <BlackList at Anitech-Systems.com>
  will be added to the BlackLists.




More information about the questions mailing list