[ndnSIM] questions regarding in-network adaptation and sliding window / congestion control implementation
poschdaniel2 at gmail.com
Wed Feb 26 04:17:13 PST 2014
I am working on the idea to use in-network adaptation (INA) in NDN. For
simulations with large numbers of hosts I want to use ndnSIM. Currently
I am struggling with two questions regarding the simulation:
1. For the transmission of multimedia data I need to request multiple
data packets at once; otherwise the transmission throughput is too low.
I have implemented a kind of "sliding-window / congestion-control
interest algorithm" in my application that extends the ndn::App class.
However, I think it might be a better choice to add this functionality
to an extended class of ndn::Face? What do you think is the correct tier
for this functionality, the ndn::Face or the ndn::App?
2. What would be the easiest way for an intermediate NDN-node to scan
the transmission channel and if necessary to modify the flowing
NDN-packets? For instance, there are three nodes connected behind one
another: dest-node <------> INA-node <------> source-node. Let's assume
that the dest-node is requesting packets of two different classes from
the source-node and the INA-node should simply drop packets of the
second class. Is it possible to implement the needed functionality for
the INA-node as ndn::App or would I have to implement this deeper in the
ndnSIM core's code? If its somehow possible I would prefer to implement
this as an extension to the ndn::App class.
Any suggestions/comments are appreciated.
More information about the ndnSIM