Additional consideration is that 1/t1 would involve divide, which will be hard 
to implement in hardware…

Regards,

Rong

On 5/26/17, 11:29 AM, "aqm on behalf of Rong Pan (ropan)" <aqm-boun...@ietf.org 
on behalf of ro...@cisco.com> wrote:

    Michael and Bob,
    
    The depart_rate is inversed in calculation delay….
    Delay = queue_length/depart_rate; 
    Hence, current_qdelay = queue_.byte_length() * PIE- 
>avg_dq_time_/DQ_THRESHOLD;
    
    Basically the average dq_time for dequeueing DQ_THRESHOLD is PIE->dq_time; 
What is the approximate time to deque the current_qlen?
    Current_qlen/DQ_THRESHOLD(what portion is current queue length relative to 
DQ_THRESHOLD)? * avg_dq_time.
    
    That is the rationale behind it.
    
    Thanks,
    
    Rong
    
        >> (EWMA) of the rate should be:
        >>
        >>      ewma(depart_rate) = DQ_THRESHOLD ∗ ewma(1/t1,1/t2,1/t3,...)
        >>                       != DQ_THRESHOLD / ewma(t1,t2,t3,...)
        >> "
        >> PIE uses the second (incorrect) formula. In the review, I discuss how
        >> wrong this could be, with an example.
        > Thanks, Bob, for pointing this out to me.
        >
        > Rong, is PIE doing this by intent (if so, what's the reason?) or is 
this
        > a flaw?
        >
    
    
    _______________________________________________
    aqm mailing list
    aqm@ietf.org
    https://www.ietf.org/mailman/listinfo/aqm
    

_______________________________________________
aqm mailing list
aqm@ietf.org
https://www.ietf.org/mailman/listinfo/aqm

Reply via email to