Analysis of fast and secure protocol based on continuous-time Markov chain

Abstract

To provide an optimal alternative to traditional Transmission Control Protocol (TCP)-based transport technologies, Aspera's Fast and Secure Protocol (FASP) is proposed as an innovative bulky data transport technology. To accurately analyse the reliability and rapidness of FASP, an automated formal technique ¿ probabilistic model checking ¿ is used for formally analysing FASP in this paper. First, FASP's transmission process is decomposed into three modules: the Sender, the Receiver and the transmission Channel. Each module is then modelled as a Continuous-Time Markov Chain (CTMC). Second, the reward structure for CTMC is introduced so that the reliability and rapidness can be specified with the Continuous-time Stochastic Logic (CSL). Finally, the probabilistic model checker, PRISM is used for analysing the impact of different parameters on the reliability and rapidness of FASP. The probability that the Sender finishes sending data and the Receiver successfully receives data is always 1, which indicates that FASP can transport data reliably. The result that FASP takes approximately 10 s to complete transferring the file of 1 G irrespective of the network configuration shows that FASP can transport data very quickly. Further, by the comparison of throughput between FASP and TCP under various latency and packet loss conditions, FASP's throughput is shown to be perfectly independent of network delays and robust to extreme packet loss.