I have to write a system that has several data processing steps, each varying in time complexity. I intend to make use of buffer queues and multithreading for each preprocessing step. I have limited number of processes (each process can be seen as a server), and memory, so the buffer queues cannot be infinite. This seems like a perfect setting to use queueing models.

However, queueing models assume markov properties for input data, as I recall they also assume the input to follow poisson distribution.

So my question is, is it okay to assume those to model the data pipeline using queueing models?

Thank you.

  • $\begingroup$ Not all of queueing theory assumes Poisson arrival process. $\endgroup$ – Mark L. Stone Jul 14 '18 at 14:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.