[ndnSIM] Measuring a download time
srk130230 at utdallas.edu
Tue Mar 1 17:01:09 PST 2016
Thank you for your help. Also, there is one more thing that I observed in my case.I have multiple consumers and producers in my scenario and for first interest packet there is a source which never sends a data back and there is another source which forwards first interest packet on other interface.
Can you please suggest me any reason for these observations?
Thanks & Regards,
From: Alex Afanasyev <aa at cs.ucla.edu>
Sent: Friday, February 26, 2016 2:37 AM
To: Komajwar, Sharvari
Cc: ndnsim at lists.cs.ucla.edu
Subject: Re: [ndnSIM] Measuring a download time
There is no simple answer to your question. Download times, besides multi/single path forwarding depends on client request pattern and congestion control (hop-by-hop and end-to-end).
ConsumerCbr may not be a great consumer to measure what you want, as it is sends interest at constant rate. If it can retrieve set of data packets over single path, it will take the same amount of time (minus reduced interest reexpressing) to fetch of multipath. You can try ConsumerWindow, but it has a very rudimentary logic that is trying to adjust client strategy to the available network capacity.
May be you should try to use different measurement parameter, which shows the effect of mutlipath strategy, without adding dependency on client requests.
> On Feb 24, 2016, at 2:12 PM, Komajwar, Sharvari <srk130230 at utdallas.edu> wrote:
> I am working on Multipath forwarding strategy for NDN. I am using ConsumerCbr application to generate interest packets. Can you please suggest me the way to measure a downloading time for a file of particular size?
More information about the ndnSIM