Why deterministic traffic shows the highest reordering ratio

In optical burst switching networks, bursts arrive out-of-sequence due to contention resolution schemes as well as contention avoidance schemes. Out-of-sequence bursts may imply out-of-sequence packets, which affect the packet layer performance. Consequently, it is necessary to classify and investigate the resulting burst/packet out-of-sequence pattern. In our previous work, we presented an analytic model to evaluate the out-of-sequence pattern for deterministic traffic showing constant inter-arrival times using the IETF WG IPPM reordering metrics. This paper extents our previous work. If the amount of out-of-order arrivals is small, the performance impact may also be small. One indication of the amount of out-of-order arrivals is the reordering ratio. In this paper, we prove analytically that deterministic traffic with constant inter-arrival time generates the highest reordering ratio for our model with respect to any other traffic characteristic. Consequently, deterministic traffic serves as an upper bound for the estimation of the reordering ratio. It allows a quick estimation of the reordering ratio and indicates if further investigations on the reordering pattern are required.