Purpose: The purpose of this master thesis is to investigate the possibility of creating a multi-thread Random Linear Network Coding Encoder which is ignorant to specific system resources, such as size of main memory and cache, and the amount of available CPU cores. This is done with the goal of significantly decreasing the latency introduced by encoding when using Random Linear Network Coding.

Method: We present three possible encoding schemes which can utilise parallel computing and multi-threaded programming, and thereby have the potential to decrease encoding latency. We select one of these schemes for implementation. With the implemented encoder we conduct an empirical study and compare the results with the latency of a state-of-the-art single threaded Random Linear Network Coding encoder.

Results: We concluded that the implemented scheme does not give a decrease to encoding latency, and that this is the effect of ignoring cache size. Additionally, we also concluded that it is unlikely, that an efficient parallel multi-threaded encoding scheme can be designed, without taking system resources, such as cache size, into account.