TCP network game: conditions for evolutionary stable equilibrium

Science China Information Sciences, 2020

TCP Vegas has been introduced in an attempt to improve the performance of TCP Reno and reduce the oscillation e ects. TCP Vegas is, however, criticized for penalizing the users using it in the presence of other users using TCP Reno. In this paper, we model the problem of congestion control as both a single shot game and a repeated game, using a game theoretic approach. We rst show that in the single shot game, the unique Nash Equilibrium Point NEP is everyone using TCP Reno, while in the in nitely repeated game there may exist an NEP that supports everyone using TCP Vegas in each period.

In this paper, we evaluate the impact of selfish behavior through a game-theoretic analysis of TCP Vegas and FAST TCP. By theoretical analysis, we obtain the Nash Equilibrium in these games and study the loss of efficiency under Nash Equilibrium. Our results show that if the users are selfish, then the efficiency of the network as a whole relies on the utility function of individual users. The efficiency of the system degrades under some conditions and remains optimal under other conditions. This motivates us to borrow the notions of Price of Anarchy and Tragedy of the Commons in Game Theory to describe this efficiency loss. We calculate the Price of Anarchy and provide the necessary and sufficient condition under which the TCP Vegas Game (FAST TCP Game) is a Tragedy of the Commons. Finally, we validate the theoretical results by the simulations carried out using NS-2.

2000

We consider network congestion problems between TCP flows and define a new game, the Window-game, which models the problems of network congestion caused by the competing flows. Analytical and experimental results show the relevance of the Window-game to real TCP congestion games and provide interesting insight into the respective Nash equilibria. Furthermore, we propose a new algorithmic queue mechanism, called Prince, which at congestion makes a scapegoat of the most greedy flow. We provide evidence which shows that Prince achieves efficient Nash equilibria while requiring only limited computational resources.

Operations Research, 2007

We consider congestion pricing as a mechanism for sharing bandwidth in communication networks, and model the interaction among the users as a game. We propose a decentralized algorithm for the users that is based on the history of the price process, where user response to congestion prices is analogous to “fictitious play” in game theory, and show that this results in convergence to the unique Wardrop equilibrium. We further show that the Wardrop equilibrium coincides with the welfare-maximizing capacity allocation.

Proceedings of the fifteenth annual ACM symposium on Parallel algorithms and architectures - SPAA '03, 2003

The well-known Transport Control Protocol (TCP) is a crucial component of the TCP/IP architecture on which the Internet is built, and is a de facto standard for reliable communication on the Internet. At the heart of the TCP protocol is its congestion control algorithm. While most practitioners believe that TCP congestion control algorithm performs very well, a complete analysis of the congestion control algorithm is yet to be done. A lot of effort has, therefore, gone into the evaluation of different performance metrics like throughput and average latency under TCP. In this paper, we approach the problem from a different perspective and use the the competitive analysis framework to provide some answers to the question "how good is the TCP/IP congestion control algorithm?" First, we prove that for networks with a single bottleneck (or point of congestion), TCP is competitive to the optimal centralized (global) algorithm in minimizing the user-perceived latency or flow time of the sessions, provided we limit the adversary by giving it strictly less resources than TCP. Specifically, we show that with O(1) times as much bandwidth and O(1) extra time per job, TCP is O(1)-competitive against an optimal global algorithm. We motivate the need for allowing TCP to have extra resources by observing that existing lower bounds for nonclairvoyant scheduling algorithms imply that no online, distributed, non-clairvoyant algorithm can be competitive with an optimal offline algorithm if both algorithms were given the same resources. Second, we show that TCP is fair by proving that it converges quickly to allocations where every session gets its fair share of network bandwidth.

2006

We consider congestion pricing as a mechanism for sharing bandwidth in communication networks, and model the interaction among the users as a game. We propose a decentralized algorithm for the users that is based on the history of the price process, where user response to congestion prices is analogous to “fictitious play” in game theory, and show that this results in convergence to the unique Wardrop equilibrium. We further show that the Wardrop equilibrium coincides with the welfare maximizing capacity allo-

European Journal of Operational Research, 2006

This paper analyzes network problems with congestion effects from a cooperative game theoretic perspective. It is shown that for network problems with convex congestion costs, the corresponding games have a non-empty core. If congestion costs are concave, then the corresponding game has not necessarily core elements, but it is derived that, contrary to the convex congestion situation, there always exist optimal tree networks. Extensions of these results to a class of relaxed network problems and associated games are derived.

Lecture Notes in Computer Science, 2004

We use the tools from game theory to understand the impacts of the inherent congestion pricing schemes in TCP Vegas as well as the problems of parameter setting of TCP Vegas on its performance. It is shown how these inherent pricing schemes result in a rate control equilibrium state that is a Nash equilibrium which is also a global optimum of the all-Vegas networks. On the other hand, if the TCP Vegas' users are assumed to be selfish in terms of setting their desired number of backlogged packets in the buffers along their paths, then the network as a whole, in certain circumstances, would operate very inefficiently. This poses a serious threat to the possible deployment of Vegas-based TCP (such as FAST TCP) in the future Internet.

2014 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), 2014

We propose a model for network optimization in a non-cooperative game setting with specific reference to the Internet connectivity. We refer to the general model shown in internal report [1], where Autonomous Systems (AS) decisions on link creation and traffic routing are strategically based on realistic interconnection costs, keeping into account the peering/transit dichotomy. Equilibria existence and convergence results were obtained in [1] only for a specific toy problem, while here we study larger scale scenarios which better fit the complex nature of the Internet. We are able to show that equilibria existence and convergence properties still hold for many possible generalizations, yet not all of them, and provide a specific example for which the system enters in a never-ending oscillation. Thanks to the use of simulations we covered those scenarios for which analytic results could not be obtained, thus analyzing a broad variety of general cases which were not studied in [1]. Simulation shows that the system, in the vast majority of cases, converges to an equilibrium. Very interestingly, even in asymmetric scenarios the equilibrium reached suggests that players tend to be symmetric with respect to the peering exchange points and send their asymmetric traffic quota via the transit service providers.