Dear ns-users, I am experimenting with ns-2 in order to understand if I can eventually use it to my project. Here is how I describe my idea:
Some mobile nodes move around a terrain and make use of the wireless network to exchange some FTP/TCP data between sources and destinations using AODV. Here is the unorthodox part: Each node will have a simple application simulated at its application layer (lets call it AppAgent). AppAgent will need to be attacked to the node and its purpose will be to do some calculations using the node positions (will query the data at some intervals) and data coming from the network layer, without making use of the network. I mean, the AppAgent will not send or receive messages through the network. Thus the AppAgent doesn't have to be attached to a transport layer agent, TCPAgent or TCPSink, or the equivalent UDP stuff. I see it as being attached to the "node" in general. So based on your experience and knowledge, is it possible to simulate an application (at the application layer of ns-2), which will do some interesting stuff by making use of data coming from lower layers of the stack, but will not require any sending/receiving of data like other applications such as Telnet for instance? My other concern is weather this simulated application will take its share of the internal clock of ns-2, as it has to run while the simulation is running and perhaps report to the trace file like the rest of the attached agents. Lets say it will have a start() command, and will start just after it is attached on its node, at the beginning of the simulation. I hope I have described the idea good enough to trigger some answers from the community. Please let me know if I don't make sense! Many thanks in advance, Alexandros
