Re: [ns] multicast support for wireless nodes

[Parveen Kumar Patel <ppatel@cs.utah.edu>]

On Thu, 19 Apr 2001, Harsh Shah wrote:

> Thanks for your replies. Actually, I cannot use two different types of
> multicast protocols for the same topology. I have to setup
> wired-cum-wireless topology and I have to use same kind of multicast
> protocol (either DM or SM) on all the nodes. 

its not a question you or me having a preference for some
protocol. traditional routing protocols do not work for ad-hoc
networks. and as you enter the wireless domain you have to have a
different protocol if you want to go over one hop from the base station. 

> As Lloyd said it might
> require several modifications in NS and I am trying to look for that. I
> have thought of doing some application level hack rather than modifying
> mobile node to support multicast. I am looking to implement virtual node
> between Base station and it's upstream router. I will enable multicast on
> only wired nodes including virtual node. My mulitcast protocol will run
> only on those wired nodes. I am using application level agent to send
> unicast packets from virtual node to downstream mobile nodes(including
> base station) when the virtual node receive mcast packets. Similarly
> the application level agent will reside on mobile nodes(receiver and
> BS) which will send joins to the upstream application level agents
> until the virtual node. Does anyone of you can think what sort of hassles
> can be there in such solution. Is it possible to have hack like this? I am
> not much acquainted with ns internals and haven't gone through its code
> yet. 

you could do that as long as you see some value in it. i don't see why
would anyone want to simulate such a thing.. primarily because it would be
hard for you to correctly simulate the application level processing. if
you really think results from such a simulation will help you in anyway,
go ahead. 

_P