[ndnSIM] Missing cached item when using a probabilisty caching

[Alex Afanasyev]

Hi Saran,

This is an artifact of policy chaining.  As I understand, the way it is chained right now, lru policy kick in first and only then the random policy.  What happens is that because LRU policy is admit always/evic when full, when CS reaches the capacity, the item is getting evicted after (to satisfy LRU policy confitions), after which processing is transferred to random policy.  Random policy does randomized accept.  Therefore, with probability (1-p) you end up with max_size-1 items.

If you chain policies the other way around, then I the behavior should be different.

---
Alex

> On Jun 24, 2014, at 11:54 PM, Saran Tarnoi <sarantarnoi at gmail.com> wrote:
> 
> Hi Alex,
> 
> Thanks for your response.
> I picked up caching probability from {0.5, 1.0, 0.01} and used it along with Lru policy. The simulation time was 100,000 sec with 10 Hz of request rate. 
> All cases showed the same problem, that is the number of items simultaneously residing in CS = CS size - 1 most of the time during simulations.
> 
> I observed that right after the item has been cached in CS, the number of items simultaneously residing in CS = CS size. 
> However, when a successive content object arrives at the node, one of items in the CS is erased, even though this content object is not cached, considering 
> ns3_rand_.GetValue () < probability_ in probability-policy.cc is now false.
> 
> I am trying to under stand how you chain the probability-policy.cc, lru-policy.cc, and trie-with-policy.cc and hope that I will understand more.
> 
> Thank so much for your time.
> 
> --
> Saran
> 
> 
> 
> 
> 
> 
> 2014-06-25 15:35 GMT+09:00 Alex Afanasyev <alexander.afanasyev at ucla.edu>:
>> Hi Saran,
>> 
>> What probability did you set and how long was your experiment? Could that be a factor?
>> 
>> ---
>> Alex
>> 
>> > On Jun 24, 2014, at 10:21 PM, Saran Tarnoi <sarantarnoi at gmail.com> wrote:
>> >
>> > Hello Alex and All,
>> >
>> > When I use probability caching, I faced a wired thing when I checked the items cached in Content Store (CS). I found that the number of items actually cached in the CS is less than the determined CS size. This wired thing disappeared when the caching probability is 1.0. Could you give me an idea why this is happening?
>> >
>> > Note that I checked this by adding
>> > m_contentStore->Print (std::cout);
>> > in
>> > ForwardingStrategy::OnInterest (Ptr<Face> inFace, Ptr<Interest> interest)
>> > of ndn-forwarding-strategy.cc.
>> >
>> > Thank you so much for your time.
>> >
>> > --
>> > Regards,
>> > Saran Tarnoi
>> > _______________________________________________
>> > ndnSIM mailing list
>> > ndnSIM at lists.cs.ucla.edu
>> > http://www.lists.cs.ucla.edu/mailman/listinfo/ndnsim
> 
> 
> 
> -- 
> Regards,
> Saran Tarnoi
> Graduate Student
> Department of Informatics
> The Graduate University for Advanced Studies (Sokendai)
> Tokyo, Japan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/ndnsim/attachments/20140625/cfa5bf60/attachment.html>