[ndnSIM] Question about the inconsistent entry in PIT
Alex Afanasyev
alexander.afanasyev at ucla.edu
Tue Sep 10 10:21:37 PDT 2013
Hi Saran,
Can you try adding this line
interest->SetInterestLifetime (m_interestLifeTime);
into apps/ndn-consumer-zipf-mandelbrot.cc after line 184.
I have a suspicion that this could be the reason of the strange results (and I really wonder how it was working in the first place without interest lifetime to be set in this consumer...)
---
Alex
On Sep 10, 2013, at 7:22 AM, Saran Tarnoi <sarantarnoi at gmail.com> wrote:
> Dear Alex,
>
> I was about to modify ndn-forwarding-strategy.cc and I found an ambiguous thing.
>
> In
> ForwardingStrategy::OnData (from line 224,)
> you put some line of codes for the case that pitEntry == 0,
> where Ptr<pit::Entry> pitEntry = m_pit->Lookup (*data);.
>
> I ran simulation in a fixed topology with ZipfMandelbrot consumer app.
> Therefore, a data packet should always follow the path used by it's interest packet.
>
> I found that if I set the frequency of request rate to 1, there was no single time that pitEntry = 0. Everything seemed to work as it should do.
>
> In contrast, if the request rate was set to 10 or greater values, pitEntry was sometimes equal to 0 at some nodes.
> In other words, the considering data packet was logically unsolicited, but in fact there should be a pending interest waiting for that data at this node.
>
> I really don't understand the reason behind this case.
> I have checked the PIT's size of those nodes, it was either 0 or 1, so it did not seem to be the problem of PIT size.
> I attached my scenario and topology file to this email.
> If you have some time, please help me explain this.
>
> Thanks so much for your help.
>
> Regards,
> Saran Tarnoi
>
> <ndn-intersection.cc><topology_intersection>
More information about the ndnSIM
mailing list