Despite the existence of several metrics to perform measurements on out-of-order packets, few works have used these metrics for comparative purposes. A potential reason for this is that the use of these simple singleton metrics makes it difficult to analyze all the effects of packet reordering. On the other hand, more complete metrics are represented in a vectorial manner, making comparative analysis challenging. In this paper, we present a scenario for testing and describe a methodology for conducting experiments to compare network paths with respect to unordered packets. The results of several simulations explore simple packet reordering metrics derived from vector metric that may allow future work to be benchmarked against. We demonstrated the behaviour of some TCP congestion control algorithms by adjusting different levels of reordering. We highlight good results with the Entropy reorder metric.
展开▼
