[ndnSIM] Caching probability issue

[G.Sánchez]

The problem was already solved. The answer was the following:

It is because of the consumer application,

// Consumer
ndn::AppHelper consumerHelper ("ns3::ndn::ConsumerCbr");.

The "ns3::ndn::ConsumerCbr" generates a series of "sequential interest
packets"; for example, Consumer will request /prefix/0, /prefix/1, ....

Therefore, the Consumer will never request any previously requested content
object, so CacheHit is always zero.
By contrast, there is more than one Consumer in tree topology. As a result,
a content object that has requested by a Consumer is also requested by
another Consumer, so CacheHit occurs.

To see what is actually generated by Consumer while running a simulation,
you can put the command below to see the log,

NS_LOG=ndn.Consumer ./waf --run=ndn-simple

Thanks to all.




2014-06-03 13:37 GMT+09:00 G.Sánchez <g.sanchez148 at gmail.com>:

> Hi all,
>
> Firstly, I'm quite new with ndnSim, so sorry if my questions may seem too
> obvious.
>
> I am trying to implement scenarios with caching probability, i.e. each
> node has a probability X to cache the content or not. As far as I know,
> this is done by using the following function inside the .cc scenario:
>
>
>
> *ndnHelper.SetContentStore ("ns3::ndn::cs::Probability::Lru",
> "CacheProbability", "1.0");*
> before installing the ndnHelper on the nodes.
>
> I have tried it with 2 scenarios, the tree scenario with cs tracers, and
> the ndn simple scenario (adding the same cs tracer). Then I check the cs
> tracer output (CacheHits/Misses), to see if the behaviour is as expected.
>
> In the tree scenario, it works correctly, depending on the probability, we
> have a number or another of CacheHits, but in the ndn simple scenario,
> everything is missed (no CacheHits), even if probability is set to 1.0,
> offering the same results no matter which probability is used. The rest of
> the code has remained untouched.
>
> Does the simple scenario have something special that the tree one doesn't
> for this situation? Should I add something else in the code? Something
> different? Is it a bug?
>
> Thank you all for your time and help.
>
> Guillermo Sánchez
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/ndnsim/attachments/20140624/92bfd1da/attachment.html>