This year's International Workshop on OMNeT++ was held on March 19, 2010
in Torremolinos, in conjunction with SimuTools 2010 as usual. On behalf of the
Program and Organizing Committee we can say that the workshop was a great success.
We had about 10 published papers and more than 30 participants in the room.The workshop programme (including presentation slides) can be read on the workshop's web site.
Five papers have been published as regular papers with an oral presentation. Thanks to the presentation slots of 35 minutes, a fruitful discussion was started after each presentation, as the authors had the possibility to present a detailed insight of their paper. The other five papers have been published as posters also including a short presentation slot of 15 minutes. We also had the possibility to discuss community related topics. Alfonso Ariza Quintana raised the idea of modularizing the INET Framework in his presentation. Andras Varga presented about the new signals feature of OMNeT++ and about extending the IDE with wizards. The participants attested a very high quality of presentations. Thanks to the strong TPC, we have had four reviews per paper and a rebuttal period, which helped to select high quality papers for the workshop. Our general impressions was that this community is advancing and growing very fast, and with good elements to work with.
We discussed several topics in the Panel Session...
Feedback about long presentation slots was generally positive. It was agreed that we want to continue the tradition of annual workshops.
The question was raised on how to aid students to start with OMNeT++, since there is evidence that a lots of newcomers on the community are students. It was proposed that we should write a book about M&S of communication networks in OMNeT++. We agreed that the first step should be to write a comprehensive manual for the INET Framework and INETMANET, possibly also MiXiM, Castalia, etc., and later on this manual could be edited into a book.
Another question was how to improve the popularity and acceptance of OMNeT++ in the network simulation field. We identified credibility as a key factor that needs to be improved. Credibility can be achieved via more thorough validation and verification of models, which can only be achieved if the development of INET/INETMANET is strengthened. The development needs to become more coordinated, with people involved in identifying tasks, reviewing code and test cases, and so on. The following steps were identified:
- open a new mailing list for model developers to make communication and cooperation between contributors easier;
- introduce nightly builds and regression tests to ensure stability; results of each build should be automatically published on a web site;
- set up requirements for "stable" models (documentation, validation tests, regression tests, code style, maintainer, etc); models that do not meet these requirements should be clearly labelled as such;
- patches to stable models should be reviewed by others (e.g. by at least one person different from the submitter) before being applied
- stable releases (probably time-based releases like Ubuntu, because INET/INETMANET consists of many independently developed models which are difficult to synchronized);
- on the long term, INET/INETMANET may be split up to several components, with a clear dependency tree; however it is advisable to keep them in a single (git) repository.
In conclusion, we see that people is actively discussing about modeling and simulation based on OMNeT++. We evidenced people is starting to work together (and publishing together) as results of the interaction in this community. These signs are telling us that we are advancing in the good direction. Also we identified some new active members that, in the future, could be very valuable respecting their contributions.
At the community dinner in the evening we decided to meet again next year for the fourth international workshop on OMNeT++.