[ndnSIM] Obtaining Metrics correctly

Md Ashiqur Rahman marahman at email.arizona.edu
Wed Jul 18 16:10:10 PDT 2018


Hello All,

I have a simple wireless ad-hoc scenario with custom flooding strategy
(each node has only one interface) as follows:

(C1)
 |  \
 |   \
 |   (A)---(B)---(P)
 |   /
 |  /
(C2)

​Where wireless reachability from one node to another is represented by the
​connecting lines. C1 and C2 are two consumers and P is the producer while
A and B are two forwarder nodes. C1 and C2 send out a single Interest with
same data name and ContentStore is enabled everywhere. Here are the steps
observed in simulation:


   1. C1 broadcasts new Interest (C2 and A receives, updates PIT entries)
   2. C2 and A forwards Interest
      - From C2 -> C1 and A receives duplicate Interest, drops (PIT
      unchanged)
      - From A -> C1, C2 and B receives, duplicate at C1, C2, drops. B gets
      as new Interest, update PIT.
   3. B forwards Interest (A receives duplicate, drop. P receives new
   Interest, ready to send data)
   4. P broadcast Data (B receives, put in CS, update PIT).
   5. B forwards Data (A and P receive, A puts in CS.. P drops).
   6. A forwards Data (C1 and C2 receive, C1 has data satisfied, C2 has PIT
   entry from C1 satisfied, both put in their CS).
   7. C2 forwards data (C1 drops, already satisfied)
   8. C2 broadcasts new Interest (same data asked by C1)
      - Has data already in CS, satisfied.


​In this example, when collecting Metrics, C2 has InSatisfiedInterests=0 even
though data is satisfied from its own cache. Which field in the metrics
should I use to correctly collect number of satisfied Interests issued by a
particular Consumer in this scenario? In case of wired scenarios, I used
the InSatisfiedInterest filed.

Best-
Ashiq

-- 
Md Ashiqur Rahman
Research Assistant, Network Research Lab
Dept. of Computer Science
The University of Arizona
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/ndnsim/attachments/20180718/c0520339/attachment.html>


More information about the ndnSIM mailing list