Hi back, I had a look at your questionnaire and I think it is too narrow. It really always depends on what you want to analyse and show. If you want to get an early insight in behaviour, a round based simulator might be OK. Yet, if you want to really understand the system (and publish..), especially if you want to understand anything with respect to time, robustness, scalability, cost, you absolutely /must/ use a discrete event packet level simulation. They of course don't scale to multiple millions of nodes (you may be lucky if you're able to simulate over 100k nodes in that case), but all other results are not meaningful.. If, on the other hand you plan to simply find out the resistance of the system towards attacks (e.g., due to topological reasons), a very simple straight forward round based analysis might be sufficient. If you want to work with undergrads you might not want to use C/C++ since then you may not trust the results, if you're a PhD student and want to come up with results meaningful enough for your thesis, you absolutely need the scalability of C/C++ as compared to Java or <you-name-it> (in my opinion, of course...)! I personally don't think there can/should be a swiss army knife for P2P simulations, since the tool will always depend on the task, and the tasks vary dramatically... ;)
just my 2ct... Thorsten Ah, and as a PS: whenever using somebody else's simulator make sure you check and /really/ understand their code. We've been looking into plenty of the existing packages and were quite appalled at times (not that they were necessarily useless in indicating some properties that the original authors had intended to use them for, but don't just use the simulator from down the street and guess it's gonna give you what you think). On 03/11/2011 02:05 PM, Simon Fleming wrote: > Hello, > > We are writing a survey paper for ACM CSUR on peer-to-peer simulators. > We are emailing this mailing list as we are interested in your views and > opinions on an ideal p2p simulator so we have made a tiny questionnaire. > > We would greatly appreciate if you spent 2 minutes filling this in as we > hope to use it to find out what researchers would want in an ideal p2p > simulator. > > Please find the questionnaire at the following URL: > > http://goo.gl/zTCKO > > All answers are anonymous, so feel free to be as critical as you like! > > Thank you for your time and help with this research. > > Best Regards > Foundations of Software Systems Group, University of Sussex. > > > > _______________________________________________ > p2p-hackers mailing list > p2p-hackers@lists.zooko.com > http://lists.zooko.com/mailman/listinfo/p2p-hackers _______________________________________________ p2p-hackers mailing list p2p-hackers@lists.zooko.com http://lists.zooko.com/mailman/listinfo/p2p-hackers
