I am considering strategies to compare the performance of routers with buffers vs. their lower latency counterparts in a simple network setting (N clients communicating with each other randomly through a router). I know that buffered routers should ideally perform better when there are traffic bursts.

My question is regarding my concern which is that I do not know what is a good strategy to pick scenarios and how could I convey that these models are relevant to common real world situations. So the question would be:

  1. How to pick scenarios that are similar to real world situations?
  2. How to convince that these scenarios are indeed relevant to real world situations.

My main concern is measuring the benefit of buffers vs. an improvement in RRT (lower latency) with varying levels of congestion. Also showing possible disadvantages where the amounts of congestion are overwhelming (buffer bloat) and possibly other patterns.


closed as unclear what you're asking by D.W., David Richerby, Luke Mathieson, vonbrand, Wandering Logic Apr 1 '14 at 22:04

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ this sounds like Network Engineering unless you are doing simulations ... are you working with real hardware and wanting to do hardware benchmarking? $\endgroup$ – vzn Feb 20 '14 at 16:04
  • 1
    $\begingroup$ So what is your question, exactly? I don't see a question in here. (If you are asking "What simulated workloads should I use?", the answer is going to depend heavily on what you are trying to accomplish, what deployment scenarios you are targeting, and other details -- I am not optimistic that we can give you a one-size-fits-all answer.) $\endgroup$ – D.W. Feb 21 '14 at 0:09
  • $\begingroup$ I am working with ns3 and doing software simulations. @D.W. I will try to modify the question to better pinpoint the area I would like to shed some light upon. $\endgroup$ – wolfdawn Feb 21 '14 at 10:14