RNC Node Level Regression test of Robustness Suite

Abstract: The main objective of this master thesis is to reconstruct randomized rendered scenarios leading to failure in the 3G system. The incitement is the ability to easily rerun the scenario triggering a failure, take additional traces to isolate the root cause and verify corrections. This is achieved by utilizing CPPemu (Cello Packet Platform) which emulates a complete 3G system, consisting of the RNC (Radio Network Controller), CN (Core Network), RBS (Radio Base Station), UE (User Equipment), where the three latter are simulated by 3Gsim. A great variety of errors were discovered during the thesis in the simulated environment. This affected the ability to run traffic unimpededly and thus the quest of reconstructing the randomized rendered scenario. One of the major problems is that 3Gsim generates incorrect mobility events when run with random mobility. Another problem is that the time between the executing event are very tight, which requires an unreasonable high precision of the timing in the environment. Approximately ten percent of the scenarios were successfully reconstructed. Although the success rate of the test runs were widely spread. In order to tackle the problems a lot of timing parameters could be tried out, or the more realistic solution is to modify the behavior of the environment itself to prevent unacceptable timings in the random mobility.