Congestion Control Dependability Assessment

We have examined the use of fault injections to evaluate the dependability of a transport layer protocol (i.e., TCP) in wireless sensor networks (WSN). We have focused on the layer's service i.e. congestion control, then, we have defined workload, faultload and dependability measures. The workload has two main components which are: the WSN architecture and the execution profile. We have defined in this paper the fault that may face different services and present their implementations in order to specify the faultload. We have introduced two dependability measures which are tolerance threshold and recovery time. We have demonstrated that the communication is interrupted after exceeding a given tolerance threshold (below this limit, the service employs a recovery time to resume execution). Our benchmarking targets eight different congestion control algorithms and presents several experiments to assess their dependability. According to our measures, BIC is the most dependable algorithm whereas Newreno and veno are the least dependable.