Context replication is a new context initialization technique for header compression proposed by the IETF. It increases header compression gains by reducing the size of bulky initialization packets. Context replication compresses header fields based on the redundancy between flows, rather than within flows. For context replication to be performed efficiently, an elaborate understanding of inter-flow field behavior is required. However, inter-flow field behavior is more complicated and relatively unstudied compared to well-known intra-flow field behavior. Also, because context replication is a new technique, much is still unknown about the range of efficiency gains achievable. In this paper, we present a novel approach to obtain and analyze inter-flow field behavior. Our approach allows the optimization of encoding specifications in context replication. By implementation, we show that our optimized encodings produce leaner initialization packets compared to current proposals. We also obtain from implementation useful insights on the range of compressed header sizes and efficiency gains in context replication.
