Costas Courcoubetis, Vasilios A. Siris, and George D. Stamoulis. Integration of pricing and flow control for ABR services in ATM networks. Proceedings of Globecom'96, November 1996.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....also fully distributed (scalable) is robust, adapts quickly to changes, has low overhead, and results in efficient resource usage. Our work is based on results from microeconomics [6] There have been other proposals to apply microeconomics to allocation of network resources, including [7] 8] [9] [10] 11] The differences of our work with those include: Our method regards fairness as a policy; we present a mechanism that can support a variety of fairness goals, or policies. Our method is fully distributed and has been shown to converge quickly under dynamic and realistic network ....

C. Courcoubetis et at., Integration of Pricing and Flow Control for ABR Service in ATM Networks, Proc. of GLOBECOMM, IEEE, 1996, pp. 644-468.

....is to achieve stability, fairness and robustness. Ideally one should design the link and source algorithm jointly so that they work in concert to steer the network to track a possibly moving desirable operating point. This motivates a recent approach to flow control based on optimization e.g. [6], 14] 16] 12] 18] 23] 19] 1] 13] 25] 26] 17] 3] where the goal is to choose source rates to maximize a global measure of network performance. Flow control, both the link and the source algorithms, is derived as a distributed solution to this welfare maximization problem. ....

Costas Courcoubetis, Vasilios A. Siris, and George D. Stamoulis. Integration of pricing and flow control for ABR services in ATM networks. Proceedings of Globecom'96, November 1996.

....change slowly compared with convergence time. There is a tremendous literature on flow control, including early schemes based on practical experience, e.g. 14,11,26] and recent schemes based on control theory, e.g. 3,6,27,8] Optimization based flow control have been proposed in, e.g. [12,13,9,15,16,18,22]. All these works motivate flow control by an optimization problem and derive their control mechanisms as solutions to the optimization problem. They differ in their choice of objective functions or their solution approaches, and result in rather different flow control mechanisms to be implemented ....

Costas Courcoubetis, Vasilios A. Siris, and George D. Stamoulis. Integration of pricing and flow control for ABR services in ATM networks. Proceedings of Globecom'96, November 1996.

....sharing the same bottleneck links should transmit, e.g. 1, 21] Such schemes require multiple bits per feedback. A more important disadvantage, however, is that the control decision is made inside the network without regard to different desires of various sources. A third type of recent works [7, 9, 5, 11, 12, 14]. motivate flow control by an optimization problem and derive their control mechanisms as a solution to the optimization problem. These papers differ in their choice of objective functions or their solution approaches, and result in rather different flow control mechanisms to be implemented at the ....

....and the network links. In [7, 9] the goal is to minimize a cost function of source rates using gradient type algorithm. A key difference between this work and ours is our solution through the dual problem and the resulting simple algorithm. Though in different form the social welfare function in [5] is actually a special case of our formulation here with a specific utility function. Their algorithm can also be considered as scaled gradient projection algorithm. They however do not provide any proof of convergence. Our model is closest to that in [11, 12] There however the overall objective ....

Costas Courcoubetis, Vasilios A. Siris, and George D. Stamoulis. Integration of pricing and flow control for ABR services in ATM networks. Proceedings of Globecom'96, November 1996. 12

....update, but now also projects the new price onto the hyperplane in Theorem 6. We illustrate through an example that this might significantly speed up the convergence towards an equilibrium. 2. 3 Related works Pricing network services and resources have received much attention recently; see e.g. [3, 14, 20, 17, 11, 4, 12][31, Chapter 8] and references therein. This paper differs from the previous work in three respects. First we consider both bandwidth and buffer in our model. Second, and more importantly, we consider both variable and fixed resources where users must balance the higher expected return of variable ....

Costas Courcoubetis, Vasilios A. Siris, and George D. Stamoulis. Integration of pricing and flow control for ABR services in ATM networks. Proceedings of Globecom'96, November 1996.

.... change slowly compared with convergence time [15] There is a tremendous literature on flow control, including early schemes based on practical experience, e.g. 11, 7, 22] and recent schemes based on control theory, e.g. 1, 4, 23, 5] Optimization based flow control have been proposed in [8, 10, 6, 13, 14, 16, 15, 18]. All these works motivate flow control by an optimization problem and derive their control mechanisms as solutions to the optimization problem. They differ in their choice of objective functions or their solution approaches, and result in rather different flow control mechanisms to be implemented ....

....measured by bandwidth prices, is determined by the interaction of sources through the capacity constraints. Another important difference is our solution of the optimization through the dual problem and the resulting simple switch algorithm. Though in different form the social welfare function in [6] is actually a special case of our formulation here with a specific utility function. Their algorithm can also be considered as a scaled gradient projection algorithm for the dual problem. Our model is closest to that in [13, 14] There however the overall objective of maximizing the total utility ....

Costas Courcoubetis, Vasilios A. Siris, and George D. Stamoulis. Integration of pricing and flow control for ABR services in ATM networks. Proceedings of Globecom'96, November 1996.

....and Reno) has been in widespread deployment. It is thus unclear how well RED and Reno will work together. Indeed the behavior of a network of RED routers shared by a set of Reno sources can be difficult to understand. This motivates a recent approach to flow control based on optimization e.g. [5], 16] 18] 13] 20] 23] 21] 1] 14] 25] 26] 19] 2] where the goal is to choose source rates to maximize a global measure of network performance. Flow control, both the link and the source algorithms, is derived as a distributed solution to this welfare maximization problem. ....

Costas Courcoubetis, Vasilios A. Siris, and George D. Stamoulis. Integration of pricing and flow control for ABR services in ATM networks. Proceedings of Globecom'96, November 1996.

....original scheme but enjoys a much better convergence property. This paper provides another example on how the optimization framework can be exploited to systematically refine REM. There is a tremendous literature on flow control. The works closest to this paper are optimization based [8] 9] [10], 11] 12] 13] 2] 5] 1] where the problem is formulated as one of optimizing a social welfare and the flow control mechanisms are derived as solutions to the optimization problem. They differ in their choice of objective functions or their solution approaches, and result in rather ....

Costas Courcoubetis, Vasilios A. Siris, and George D. Stamoulis, "Integration of pricing and flow control for ABR services in ATM networks," Proceedings of Globecom'96, November 1996.

....TCP flow control [15] and recent enhancement in [10] the binary feedback schemes of, e.g. 29] 6] two bit feedback scheme of [23] the control theoretic approach of, e.g. 3] 30] 7] etc. Also see a recent review in [14] A key premise of optimization based flow control [11] 13] [8], 16] 17] 12] 20] 19] 26] 22] is that sources with different valuation of bandwidth should react differently to network congestion. All these works motivate flow control by an optimization problem and derive their control mechanisms as solutions to the optimization problem. They ....

Costas Courcoubetis, Vasilios A. Siris, and George D. Stamoulis. Integration of pricing and flow control for ABR services in ATM networks. Proceedings of Globecom'96, November 1996.

No context found.

Costas Courcoubetis, Vasilios A. Siris, and George D. Stamoulis. Integration of pricing and flow control for ABR services in ATM networks. Proceedings of Globecom'96, November 1996.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC