Why two languages? uses two languages because simulator has two different kinds of things it needs to do. On one hand, detailed simulations of protocols requires a systems programming language which can efficiently manipulate bytes, packet headers, and implement algorithms that run over large data sets. For these tasks run-time speed is important and turn-around time (run simulation, find bug, fix bug, recompile, re-run) is less important.
On the other hand, a large part of network research involves slightly varying parameters or configurations, or quickly exploring a number of scenarios. In these cases, iteration time (change the model and re-run) is more important. Since configuration runs once (at the beginning of the simulation), run-time of this part of the task is less important.
meets both of these needs with two languages, C++ and OTcl. C++ is fast to run but slower to change, making it suitable for detailed protocol implementation. OTcl runs much slower but can be changed very quickly (and interactively), making it ideal for simulation configuration. (via tclcl) provides glue to make objects and variables appear on both langauges.
For more information about the idea of scripting languages and split-language programming, see Ousterhout's article in IEEE Computer [#!Ousterhout98a!#]. For more information about split level programming for network simulation, see the ns paper [#!Bajaj99a!#].
Which language for what? Having two languages raises the question of which language should be used for what purpose.
Our basic advice is to use OTcl:
and use C++:
For example, links are OTcl objects that assemble delay, queueing, and possibly loss modules. If your experiment can be done with those pieces, great. If instead you want do something fancier (a special queueing dicipline or model of loss), then you'll need a new C++ object.
There are certainly grey areas in this spectrum: most routing is done in OTcl (although the core Dijkstra algorithm is in C++). We've had HTTP simulations where each flow was started in OTcl and per-packet processing was all in C++. This approache worked OK until we had 100s of flows starting per second of simulated time. In general, if you're ever having to invoke Tcl many times per second, you problably should move that code to C++.