Hi, 

I have been following the email thread the
synchronization issue for UDP CBR flows.

http://mailman.isi.edu/pipermail/ns-users/2006-June/056044.html

I am curious in representing the drop-tail queue with
finite buffer with Kendall notation. Since it is UDP
CBR packets and FIFO scheduling. From my
understanding, it can presented as d/d/1/k-1 where d
representing deterministic interarrival and service
time. Finally, k-1 represents the finite buffer.

In the email thread, it mentioned in order to
eliminate the synchronization issue between flows and
closely emulate real world packet transmission from
the traffic source. Randomness has to be introduced by
setting random_ to true: "$cbr_name set random_ 1".

My Question is, does d/d/1/k-1 is a suitable kendall
notation when the random value from uniform
distribution is included in calculating the CBR
interarrival time ( i.e. t =  delay + delay *
uniform(-0.5,0.5) )? Can you recommend reference which
has the formula for Items waiting, waiting time, Items
residing and Resident time?

Honestly, I'm lack understanding of the queueing
theory. Currently reading a book by  Thomas G.
Robertazzi. Please give some hints where I can find
the information on d/d/1/k for droptail. Thank you

Ben.

Send instant messages to your online friends http://uk.messenger.yahoo.com 

Reply via email to