ICN Simulator on omnetpp

[Paul S]

I'm starting this as a discussion as I expect to be asking several stuff here over time. I'm really new to omnetpp. Really new, and I have to make an ICN network simulator.
 I read almost half the manual and I want to confirm some of my thoughts about the implementation I have to do.

I was thinking:
1) Initially I don't incorporate INET at all. I just build each smart node of the network, connect them with channels that have a specific rate, delay etc and they handle the forwarding, publish/subscribing etc. When that part is complete I can possibly (if more realistic simulations are needed) add INET as well and make each node communicate as it did before but over ethernet/wifi/ppp (so layer 2).
How feasible is that? Will I just waste my time later on trying to do it like that?

2) From what I understand I can easily define specifically structured packets in omnet and use them in my nodes? What restrictions do apply in those conserning size?
If I later on want to simulate video/sound transfer and use larger frames can I do that with ease? How will that be extended if I decide to add INET as I describe above?

3) I'm thinking that each node will have lots of different modules inside (randevou module, forwarding module, possibly topology module) and it will also provide an api for higher level services that want to communicate on ICN networks. All these modules will be connected with simple input/output channels and will send specifically structured messages between them. That way it will be easy to implement different algorithms and swap modules to test them. Is that feasible? Would you recommend something specific for this?

All other help is appreciated. :)
Kind regards.

[Michael Kirsche]

Hmm by ICN you mean Information-Centric Networks?

If so, there is a ICN simulator based on OMNeT and INET by Vastardis Nikolaos available: http://privatewww.essex.ac.uk/~nvasta/ICNSim.htm

[P S]

Yeap that's what I meant.
I hadn't seen the simulator before. I'll check it out but I'll
probably still have to either implement it separately or improve it.
So the questions still stand even out of pure interest. :)

Thank you.

> --
> You received this message because you are subscribed to a topic in the
> Google Groups "omnetpp" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/omnetpp/BWjqx4S3TAo/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to
> omnetpp+u...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.

[Paul S]

I checked it out a bit and it's based on blackadder which I won't be using.
My goal is to basically implement a different architecture not utilize an existing one.

[Michael Kirsche]

Check the simulator, if it provides what you need, stick with it.
If not, check if you can extend it.
If you need to extend it, you will have to stick with INET.

A complete redesign and redevelopment of such a complex simulator will prove to be quite complex, so my advice would be to stick with existing simulators (if available) or existing frameworks (if no simulator is available) and build on-top instead of refactoring your model/framework during the development. Introducing the INET modules later during a development will make many developments that you might have done before (without INET) obsolete.

Either way, analyze the ICN simulator beforehand and try it.

[Manolis]

If you have interest to use the CCNx arch for ICN simulations, then

1) there is the ccn-lite project at www.ccn-lite.net. It is a light-weight implementation of CCN. The same codebase can run in kernel space, user space, and within OMNET++ !
2) there is ndnSim, which is an implementation of CCNx for ns2/3

best

M.