RoutingSim

RoutingSim has been designed as a scalable simulation environment for the analysis of routing protocols in potentially dynamic environments where nodes and links of the topology may fail.

• RoutingSim builds on the discrete event simulator ​OMNeT++ and uses its simulation kernel as a solid foundation to build on.
• By focussing on the network layer, RoutingSim achieves a high scalability and has been used to perform simulations with up to 200000 nodes.
• To enable the analysis of routing protocols in dynamic environments where nodes and links may fail, RoutingSim takes control of the topology management
  • RoutingSim can use the OMNeT++ NED language to define network topologies but is also able to use topology files as input and to this purpose supports various file formats.
  • To investigate routing protocols on dynamic topologies, RoutingSim offers various ready-to-use dynamic models that define the type, time, and duration of failures. These dynamic models can be extended in a modular fashion.
  • Failing nodes or links change the underlying topology and this needs to be reflected when determining some statistics as, e.g., the stretch. Whereas the shortest path matrix may be precalculated in static environments, this is not possible in the dynamic case but the. As oftentimes not any-to-any connectivity but only a sampling set of node pairs is examined, RoutingSim supports on-demand calculation of such metrics for only the sampled node pairs. While RoutingSim takes care of the topology management, the efficient and mature ​Boost Graph Library data structures and graph algorithms are used to, e.g., to determine connectivity or shortest paths.
  • For very large simulations it is often not practical to collect every change in a variable under observation. The RoutingSim Statistics Module optionally enables sampling of such variables at user-configurable time intervals for increased scalability.
• New routing protocols only need to implement a lean interface and can then be examined using the RoutingSim framework.