S. Chen and K. Park. An architecture for noncooperative QoS provision in many-switch systems. In IEEE INFOCOM, 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....model, and assuming a continuous range of class choices, 21] showed that when the set of user requirements is feasible, each flow converges to a class that provides the requested delay bound. Chen and Park considered individual flows with absolute performance requirements in a stateless network [22]. The class selection is formulated as an optimization problem in which the overall resource usage cost is to be minimized, subject to the constraint that the performance requirement of each flow has to be met. Interestingly, there is a distributed algorithm that solves this problem. Ren and Park ....

S. Chen and K. Park, "An Architecture for Noncooperative QoS Provision in Many-Switch Systems," in Proceedings of IEEE INFOCOM, 1999.

....and the final results are quire different. It is also likely that in practice the class prices would depend on marketing and competition, rather than on network provisioning mechanisms. Chen and Park considered individual flows with absolute performance requirements in a stateless network [30]. The class selection is formulated as an optimization problem in which the overall resource usage cost is to be minimized, subject to the constraint that the performance requirement of each flow has to be met. Interestingly, there is a distributed algorithm that solves this problem. 25 Ren and ....

S. Chen and K. Park, "An Architecture for Noncooperative QoS Provision in Many-Switch Systems," in Proceedings of IEEE INFOCOM, 1999.

....and the final results are quire different. It is also likely that in practice the class prices would depend on marketing and competition, rather than on network provisioning mechanisms. Chen and Park considered individual flows with absolute performance requirements in a stateless network [22]. The class selection is formulated as an optimization problem in which the overall resource usage cost is to be minimized, subject to the constraint that the performance requirement of each flow has to be met. Interestingly, there is a distributed algorithm that solves this problem. Ren and Park ....

S. Chen and K. Park, "An Architecture for Noncooperative QoS Provision in Many-Switch Systems," In Proceedings IEEE INFOCOM, 1999.

....work at some rate which he can change in response to charges. For example, in the model of Low and Lapsley [36] each user chooses a rate according to his preferences, and is charged, and the charges are chosen so that social welfare is maximized subject to capacity constraints. Chen and Park [6] let each user allocate his total rate among a class of services and seek to maximize social welfare, measured in terms of constraints on a fixed class of quality of service indicators such as average delay or loss. The problem with these approaches is that they ignore the random bursty nature of ....

Shaogang Chen and Kihong Park. An architecture for noncooperative QoS provision in many-switch systems. In Proceedings of IEEE Infocom, 1999. URL http://yake.ecn.purdue.edu/~shaogang.

....priority classes in order to track a socially optimal allocation. This approach is theoretically appealing, but expensive (complex) to implement. The scheme considered in this paper is simpler as prices associated with the different service classes are fixed. Park et al. also consider in [2, 16] a service where packets get marked according to customer s QoS requirements. However, there is no price associated with the different traffic classes, i.e. the costs incurred to customers are purely performance related. They study the situation where users are free to choose their traffic class ....

....with the different traffic classes, i.e. the costs incurred to customers are purely performance related. They study the situation where users are free to choose their traffic class [16] as well as where users indicate their QoS requirements and a network controller assigns network resources [2]. The priority service that we consider here can be interpreted as an approximation to the smart market proposal of MacKie Mason and Varian [11, 12] While in the smart market users can attach any bid u 2 2 . C d(1) d( i 1 i d( d(N) d( i 1 C Figure 1: The link ....

S. Chen and K. Park, "An Architecture for Noncooperative QoS Provision in Many-Switch Systems," Proc. IEEE INFOCOM '99, 1999.

....priority classes in order to track a socially optimal allocation. This approach is theoretically appealing, but expensive (complex) to implement. The scheme considered in this paper is simpler as prices associated with the different service classes are fixed. Park et al. also consider in [2, 17] a service where packets get marked according to customer s QoS requirements. However, there is no price associated with the different traffic classes, i.e. the costs incurred to customers are purely performance related. They study the situation where users are free to choose their traffic class ....

....with the different traffic classes, i.e. the costs incurred to customers are purely performance related. They study the situation where users are free to choose their traffic class [17] as well as where users indicate their QoS requirements and a network controller assigns network resources [2]. The priority service that we consider here can be interpreted as an approximation to the smart market proposal of MacKie Mason and Varian [10, 11] While in the smart market users can attach any bid u 2 to their packets, in the above priority service users are limited to choose a bid u 2 fu 1 ....

S. Chen and K. Park, "An Architecture for Noncooperative QoS Provision in ManySwitch Systems," Proc. IEEE INFOCOM '99, 1999.

....with the different priority classes in order to track a socially optimal allocation. This approach is theoretically appealing, but expensive to implement. The scheme considered in this paper is simpler as prices associated with the different service classes are fixed. Park et al. also consider in [2], 9] a service where packets get marked according to customer s QoS requirements. However, there is no price associated with the different traffic classes, i.e. the costs incurred to customers are purely performance related. They study the situation where users are free to choose their traffic ....

....with the different traffic classes, i.e. the costs incurred to customers are purely performance related. They study the situation where users are free to choose their traffic class [9] as well as where users indicate their QoS requirements and a network controller assigns network resources [2]. The rest of the paper is organized as follows. In Section II we summarize the results of [7] In Section III we present the traffic and link model that we use for bursty traffic. In Section IV, we analyze the case where users with elastic traffic access the link. In Section V, we outline how the ....

S. Chen and K. Park, "An Architecture for Noncooperative QoS Provision in Many-Switch Systems," Proc. IEEE INFOCOM '99, 1999.

....each user sends work at some rate which he can change in response to charges. For example, in the model of Low [17] each user chooses a rate according to his preferences, and is charged, and the charges are chosen so that social welfare is maximized subject to capacity constraints. Chen and Park [3] let each user allocate his total rate among a class of services and seek to maximize social welfare, measured in terms of constraints on a fixed class of quality of service indicators such as average delay or loss. The problem with these approaches is that they ignore the random bursty nature of ....

Shaogang Chen and Kihong Park. An architecture for noncooperative QoS provision in many-switch systems. In Proceedings of the IEEE Infocom, 1999. URL http://yake.ecn.purdue.edu/~shaogang.

....with the different priority classes in order to track a socially optimal allocation. This approach is theoretically appealing, but expensive to implement. The scheme considered in this paper is simpler as prices associated with the different service classes are fixed. Park et al. also consider in [1, 6] a service where packets get marked according to customer s QoS requirements. However, there is no price associated with the different traffic classes, i.e. the costs incurred to customers are purely performance related. They study the situation where users are free to choose their traffic class ....

....with the different traffic classes, i.e. the costs incurred to customers are purely performance related. They study the situation where users are free to choose their traffic class [6] as well as where users indicate their QoS requirements and a network controller assigns network resources [1]. 2 Link Model For our analysis, we assume that time is divided into slots and consider a single link with a capacity of C packets per time slot. There is no buffer available and packets that do not get transmitted in a given time slot are dropped. The link supports a finite set I = f1; Ng ....

S. Chen and K. Park, "An Architecture for Noncooperative QoS Provision in ManySwitch Systems," Proc. IEEE INFOCOM '99, 1999.

No context found.

S. Chen and K. Park, \An architecture for noncooperative QoS provision in many-switch systems," in Proc. IEEE INFOCOM '99, pp. 864-872, 1999.

....IN 47907, U.S.A. frenh,parkg cs.purdue.edu Abstract. Architecting networks capable of providing scalable, ecient, and fair services to users with diverse QoS requirements is a pressing problem. The two principal issues are: design of good per hop behavior and edge control. In previous work [2, 3], we studied aggregate ow QoS control from a noncooperative resource provisioning context. In [20] the framework was generalized by, one, solving an optimal aggregate ow per hop behavior problem, and two, showing how it can be used coupled with end to end label control to facilitate ....

....end to end. In [9] the authors describe a proportional di erentiation model which seeks to achieve robust, con gurable service class separation i.e. QoS di erentiation with the support of two candidate packet schedulers. They use simulation to study the behavioral properties. In previous work [2, 3, 18], we studied aggregate ow per hop control mechanisms and endto end controls motivated by game theoretic considerations a router performs class based label switching which emulates user optimal service class selection with respect to sel shness without considering the space of all aggregate ow ....

S. Chen and K. Park. An architecture for noncooperative QoS provision in manyswitch systems. In Proc. IEEE INFOCOM '99, pages 864-872, 1999.

....trade o relationship with each other and transporting application trac over reserved channels, in general, incurs a high cost. Aggregate ow scheduling has been investigated in the noncooperative QoS provisioning context explicitly for multi class QoS and implicitly for competitive routing. In [2, 3, 28] the authors introduced aggregate ow per hop packet scheduling mechanisms and end to end controls motivated by game theoretic considerations a router performs class based label switching which emulates user optimal service class selection with respect to sel shness without considering the space ....

.... the time complexity is O(n L) however, since the number of labels L is a small constant (at most 256 if all TOS bits are used) we treat L as a constant and write O(n) 7 This can be formally attributed in several ways, for example, using congestion pricing and utilization as cost measures [3]. 9 delivered QoS is de cient vis a vis its requirement, other things being equal, it is desirable that the lightly loaded routers take on a greater share of request for improved QoS than heavily loaded ones. We show next that both properties are satis ed by the optimal aggregate ow ....

[Article contains additional citation context not shown here]

S. Chen and K. Park. An architecture for noncooperative QoS provision in many-switch systems. In Proc. IEEE INFOCOM '99, pages 864-872, 1999.

....behavior mode intrinsic to AFEC is also the modus operandi of TCP, and from this perspective, there are no instrinsic reasons to suspect that AFEC would not become a well behaved member of the existing protocol family. A gametheoretic discussion of QoS provision issues can be found in [13,14]. In most recent work [57] we have extended AFEC to the multiple time scale trac control framework [56] a novel workload sensitive trac control paradigm where, in addition to AFEC s adaptive feedback control, long range correlation structure in self similar trac is explicitly exploited to a ect ....

S. Chen and K. Park. An architecture for noncooperative QoS provision in many-switch systems. In Proc. IEEE INFOCOM '99, pages 864-872, 1999.

....is achieved while at the same time introducing uncertainty and volatility by flowaggregation and aggregate flow packet switching per hop. B. Key Issues A number of works have studied the behavioral characteristics of specific instances of differentiated services networks. In previous work [5] [12], 13] we introduced aggregateflow per hop control mechanisms motivated by game theoretic considerations a router performs class based label switching which emulates user optimal service class selection with respect to selfish users without considering the space of all aggregate flow per hop ....

S. Chen and K. Park, "An architecture for noncooperative QoS provision in many-switch systems," in Proc. IEEE INFOCOM '99, 1999, pp. 864-- 872.

No context found.

S. Chen and K. Park. An architecture for noncooperative QoS provision in many-switch systems. In IEEE INFOCOM, 1999.

No context found.

Chen S,Park K. An architecture for noncooperative QoS provision in many-switch systems. In: IEEE INFOCOM. New York,USA,March 1999. pp. 864--72.

No context found.

S. Chen and K. Park. An architecture for noncooperative QoS provision in many-switch systems. In IEEE INFOCOM, 1999.

No context found.

S. Chen and K. Park. An architecture for noncooperative QoS provision in many-switch systems. In IEEE INFOCOM, 1999.

No context found.

Shaogang Chen and Kihong Park. An Architecture for Noncooperative QoS Provision in Many-Switch Systems. In Proceedings of IEEE INFOCOM'99, March 1999.

No context found.

Shaogang Chen and Kihong Park. An Architecture for Noncooperative QoS Provision in Many-Switch Systems. In Proc. of INFOCOM. IEEE, 2001.

No context found.

Shaogang Chen and Kihong Park. An Architecture for Noncooperative QoS Provision in Many-Switch Systems. In Proceedings of the 18th Annual Joint Conference of the IEEE Computer and Communications Societies (INFOCOM'99), pages 864--872. IEEE Computer Society Press, March 1999.

No context found.

Shaogang Chen and Kihong Park, "An Architecture for Noncooperative QoS Provision in Many-Switch Systems ," in Proceedings of the Conference on Computer Communications (IEEE Infocom) , (NY), Mar. 1999.

No context found.

S. Chen and K. Park. An architecture for noncooperative QoS provision in many-switch systems. In IEEE INFOCOM, 1999.

No context found.

S. Chen and K. Park, \An Architecture for Noncooperative QoS Provision in Many-Switch Systems," Proceedings of the IEEE INFOCOM, pp. 864-872, 1999.

No context found.

S. Chen and K. Park, \An Architecture for Noncooperative QoS Provision in Many-Switch Systems," Proceedings of the IEEE Infocom, 1999.

First 50 documents

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