Evolutionary Games in TCP Networks with Speed Restriction Policies

Competitive development of various versions of network protocols is an essential part of computer networks. The most-used protocol today is Transmission Control Protocol (TCP). There is a large number of implementations of the TCP protocol, which differ by mechanism of congestion control. TCP develops by improving its existent implementations, vanishing some of them and via creation of a new ones. The possibility of using new versions of the protocol allows the user to increase the data rate by selecting the appropriate implementation of TCP. It is difficult to predict consequences of computer network users’ interaction in situations when many users try to achieve higher data rate by applying different TCP implementation. The actual task is to develop a theoretical and program tools to model such competitive dynamic interactions. This is the goal of my scientific-research work. Game theory, which is the theory of mathematical models of optimal decision making in situations of conflicts of interest, is the best suited to solve a particular problem because it allows you to find a solution in terms of non-cooperative interaction, which usually happens between the networks TCP-connections. This paper examines the possibility of coexistence of different implementations of protocols that users can change to improve their own capacity. It also examines games between protocols in cases when users’ packets management policies are applied.