This paper studies network resource allocation between users that manage multiple connections, possibly through different routes, where each connection is subject to congestion control. We formulate a user-centric Network Utility Maximization problem that takes into account the aggregate rate a user obtains from all connections, and we propose decentralized means to achieve this fairness objective. In a first proposal, cooperative users control their number of active connections based on congestion prices from the transport layer to emulate suitable primal-dual dynamics in the aggregate rate;
We show this control achieves asymptotic convergence to the optimal user-centric allocation. For the case of noncooperative users, we show that network stability and user-centric fairness can be enforced by a utility-based admission control implemented at the network edge. We also study stability and fairness issues when routing of incoming connections is enabled at the edge router.
We obtain in this case a characterization of the stability region of loads that can be served with routing alone and a generalization of our admission control policy to ensure user-centric fairness when the stability condition is not met. The proposed algorithms are implemented at the packet level in .NET demonstrated through simulation. Network Resource Allocation for Users With Multiple Connections Fairness and Stability
Previous works in the literature typically adopt a time-scale separation assumption, which assumes that, whenever the number of users in the system changes, the data rates of the users is adjusted instantaneously to the optimal and fair rate allocation. Under this assumption, it has been shown that such rate assignment policies can achieve the largest possible stability region in this time-scale separation assumption is removed and it is shown that the largest possible stability region can still be achieved by a large class of congestion control algorithms. A second assumption often made in prior work is that the packets of a source (or user) are offered to each link along its path instantaneously, rather than passing through one queue at a time show that connection-level stability is again maintained when this assumption is removed, provided that a back-pressure scheduling algorithm is used jointly with the appropriate congestion controller
We propose a decentralized admission control rule based on user utilities and thus tailored to our proposed user-centric fairness. We analyze the performance of this control under a traffic model of random connection arrival/departures through a fluid limit argument. The mechanism is shown to protect the network from greedy users, imposing in situations of overload the desired notion of fairness.
We turn our attention to the related problem of connection-level routing: Users bring end-to-end jobs to transfer, with routes chosen by the network. While each individual connection remains single-path, users may now profit from several routes. We characterize the stability region of this problem and give conditions under which it is attainable by a simple congestion-based routing policy.
We analyzed resource allocation in networks from a connection-level perspective with the intention to bridge the gap between classical NUM applied to congestion control and the user-centric perspective. New notions of fairness appear as user utilities are evaluated on aggregates of traffic, which can model different interesting situations. We showed how the control of the number of connections can be used to impose these new notions of fairness, and how the users can cooperate in order to drive the network to a fair equilibrium.
We plan to address several theoretical and technical questions that are still open. Stability results for admission control and the stability region of the routing policy proposed are two important theoretical questions. In practical terms, it would be interesting to explore new network implementations based on current congestion notification protocols that will helpmake these decentralized admission control mechanisms scalable to large networks.