Sporadic information sources
U. Loher · 2002
Message arrivals encountered in digital transmission over most real communication channels are not independent but appear in clusters. We propose a model of such a bursty K-ary source using a Markov chain with two states. It is shown that the protocol information of this sporadic source can be drastically reduced on the one hand by not encoding intermessage information (e.g., the starting point of a packet) and on the other hand by buffering and reordering messages. Trade-offs between reduced protocol information and message delays are also considered.