N. G. Duffield and S. H. Low. The cost of quality in networks of aggregate traffic. In Proceedings of IEEE INFOCOM'98,San Francisco, March 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... results are available on statistical multiplexing of adversarial traffic, which can consider scheduling algorithms other than a simple multiplexer [7] 12] Related work to this paper are all attempts to consolidate the deterministic network calculus [4] with statistical multiplexing (e.g. 2] [6], 10] 11] 12] 14] In addition, of particular relevance to this paper are all previous results on statistical multiplexing gain with adversarial regulated traffic, as cited above. We refer to [1] for a detailed discussion of related work. The results derived in this paper only apply to a ....

N. G. Duffield and S. H. Low. The cost of quality in networks of aggregate traffic. In Proceedings of IEEE INFOCOM'98,San Francisco, March 1998.

....timescale, for any service rate larger than the mean input rate; and that this is true throughout the network; and further that in the limit the different output flows are independent. There has also been some work on the output of a router under the many sources limiting regime. Duffield and Low [18] give a large deviations principle CHAPTER 4. NETWORKS 38 for the aggregate output using the contraction principle, just as has been done for the large buffer regime. Because the many sources regime has a richer structure than the large buffer regime, it produces variational formulae that are even ....

N. G. Duffield and S. Low. The cost of quality in networks of aggregate traffic. In Proceedings of IEEE Infocom, 1998. URL http://www.research. att.com/~duffield/pubs/cost.ps.

....e Gammaz 2 =2 (6) where z and t are the same as Norros lower bound, and is a tabulated function of the Hurst parameter H . 3.4. A Large Deviations (LDV) Technique Using the theory of large deviations, one can approximate the acceptance region so that condition (3) is satisfied, as follows [3]: ffl = sup t 0 exp( Gamma A(t) Ct b) exp(sup t 0 inf s2 [stN ff(s; t) Gamma s(Ct b) or fl = Gamma sup t 0 inf s2 [stN ff(s; t) Gamma s(Ct b) 7) where A(t) is the Legendre transformation of the logarithmic moment generating function, i.e. A(t) s) ....

N.G. Duffield and S.H. Low, The cost of quality in networks of aggregate traffic, in: Proc. of the IEEE INFOCOM '98, 1998, pp. 525--532.

.... are available on statistical multiplexing of adversarial traffic, which can consider scheduling algorithms other than a simple multiplexer [8] 14] Related work to this paper are all attempts to consolidate the deterministic network calculus [4] with statistical multiplexing (e.g. 5] 2] [7], 12] 13] 14] 16] In addition, of particular relevance to this paper are all previous results on statistical multiplexing gain with adversarial regulated traffic, as cited above. We refer to [1] for a detailed discussion of related work. The results derived in this paper only apply to a ....

N. G. Duffield and S. H. Low. The cost of quality in networks of aggregate traffic. In Proceedings of IEEE INFOCOM'98, San Francisco, March 1998.

.... results are available on statistical multiplexing of adversarial traffic, which can consider scheduling algorithms other than a simple multiplexer [7] 12] Related work to this paper are all attempts to consolidate the deterministic network calculus [4] with statistical multiplexing (e.g. 2] [6], 10] 11] 12] 14] In addition, of particular relevance to this paper are all previous results on statistical multiplexing gain with adversarial regulated traffic, as cited above. We refer to [1] for a detailed discussion of related work. The results derived in this paper only apply to a ....

N. G. Duffield and S. H. Low. The cost of quality in networks of aggregate traffic. In Proceedings of IEEE INFOCOM'98,San Francisco, March 1998.

.... results are available on statistical multiplexing of adversarial traffic, which can consider scheduling algorithms other than a simple multiplexer [7] 12] Related work to this paper are all attempts to consolidate the deterministic network calculus [4] with statistical multiplexing (e.g. 2] [6], 10] 11] 12] 14] In addition, of particular relevance to this paper are all previous results on statistical multiplexing gain with adversarial regulated traffic, as cited above. We refer to [1] for a detailed discussion of related work. The results derived in this paper only apply to a ....

N. G. Duffield and S. H. Low. The cost of quality in networks of aggregate traffic. In Proceedings of IEEE INFOCOM'98, San Francisco, March 1998.

.... is interesting to note that the above approach for determining optimal token bucket parameters is related to the interpretation of the time parameter t as the ratio of the marginal cost per unit capacity over the marginal cost per unit of buffer [5] This interpretation has also been considered in [9] to guide users, or flows, to select the same ratio of token rate and bucket depth values. Such a selection makes it simpler to determine the total amount of resources required to simultaneously carry all the flows. The network and user functions are collectively shown in Table 1. As noted ....

N. G. Duffield and S. H. Low. The cost of quality in networks of aggregate traffic. In Proc. of IEEE INFOCOM'98, 1998.

.... is interesting to note that the above approach for determining optimal token bucket parameters is related to the interpretation of the time parameter t as the ratio of the marginal cost per unit capacity over the marginal cost per unit of buffer [5] This interpretation has also been considered in [8] to guide users, or flows, to select the same ratio of token rate and bucket depth values. Such a selection makes it simpler to determine the total amount of resources required to simultaneously carry all the flows. The network and user functions are collectively shown in Table 1. As noted ....

N. G. Duffield and S. H. Low. The cost of quality in networks of aggregate traffic. In Proc. of IEEE INFOCOM'98, 1998.

....these algorithms is that a user is guaranteed a minimum share of resources and gets random extra amounts depending on network condition. For elastic traffics [27] that can tolerate some degree of delay or loss, buffer is also a scarce resource to be traded off in network resource allocation, e.g. [15, 6, 23, 5]. Again buffer allocation can be implemented by schemes ranging from complete partitioning, in which all users are guaranteed fixed amounts of buffer, to complete sharing, in which no user is guaranteed any fixed amounts of buffer, e.g. 2] In this paper we describe a novel model for such ....

N. Duffield and S. Low. The cost of quality in networks of aggregate traffic. In IEEE Infocom'98, San Francisco, CA, March 1998.

....is that a user is guaranteed a minimum share of resources and gets random extra amounts depending on network condition. For elastic traffics [25] that can tolerate some degree of delay or loss, buffer is also a scarce resource to be tradedoff in network resource allocation, e.g. 11] 7] 20] [6]. Again buffer allocation can be implemented by schemes ranging from complete partitioning, in which all users are guaranteed fixed amounts of buffer, to complete sharing, in which no user is guaranteed any fixed amounts of buffer, e.g. 4] In this paper we describe a novel model for such ....

N. Duffield and S. Low. The cost of quality in networks of aggregate traffic. In IEEE Infocom'98, San Francisco, CA, March 1998.

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