## A new approach for allocating buffers and bandwidth to heterogeneous, regulated traffic in an ATM node (1995)

Venue: | IEEE Journal on Selected Areas in Communications |

Citations: | 150 - 9 self |

### BibTeX

@ARTICLE{Elwalid95anew,

author = {Anwar Elwalid and Debasis Mitra and Robert H. Wentworth},

title = {A new approach for allocating buffers and bandwidth to heterogeneous, regulated traffic in an ATM node},

journal = {IEEE Journal on Selected Areas in Communications},

year = {1995},

volume = {13},

pages = {1115--1127}

}

### Years of Citing Articles

### OpenURL

### Abstract

Abstract-A new approach to determining the admissibility of variable bit rate (VBR) traffic in buffered digital networks is developed. In this approach all traffic presented to the network is assumed to have been subjected to leaky-bucket regulation, and extremal, periodic, on-off regulated traffic is considered; the analysis is based on fluid models. Each regulated traffic stream is allocated bandwidth and buffer resources which are independent of other traffic. Bandwidth and buffer allocations are traded off in a manner optimal for an adversarial situation involving minimal knowledge of other traffic. This leads to a single-resource statistical-multiplexing problem which is solved using techniques previously used for unbuffered traffic. VBR traffic is found to be divisible into two classes, one for which statistical multiplexing is effective and one for which statistical multiplexing is ineffective in the sense that accepting small losses provides no advantage over requiring lossless performance. The boundary of the set of admissible traffic sources is examined, and is found to be sufficiently linear that an effective bandwidth can be meaningfully assigned to each VBR source, so long as only statistically-multiplexable sources are considered, or only nonstatistically-multiplexable sources are considered. If these two types of sources are intermixed, then nonlinear interactions occur and fewer sources can be admitted than a linear theory would predict. A qualitative characterization of the nonlinearities is presented. The complete analysis involves conservative approx-imations; however, admission decisions based on this work are expected to be less overly conservative than decisions based on alternative approaches. I.

### Citations

3016 |
Convergence of Probability Measures
- Billingsley
- 1968
(Show Context)
Citation Context ...ds u ji are independent, nonnegative random variables with moment generating functions M j (s) = E[exp(su ji )] = Z 1 0 e sx dW j (x) ; (3.4) where W j (x) = Pr(u jisx) : (3.5) Chernoff's bound gives =-=[2]-=-, log P losss\GammaF K (s ) ; (3.6) where F K (s) = sC \Gamma J X j=1 K j log M j (s) (3.7) and F K (s ) = sup s0 F K (s). If C !1 and K j =C = O(1), then [3], [17] log P loss = \GammaF K (s ) 1 + O `... |

1673 | A generalized processor sharing approach to flow control in integrated services networks ⎯ the single node case
- Parekh, Gallager
- 1993
(Show Context)
Citation Context ...egulators. Even with this worst case approach there are interesting choices to be made. One option is to proceed with deterministic bounds to ensure that there are no losses in the network nodes. See =-=[16]-=- for an example along these lines. Another option is to allow very small loss probabilities, say 10 \Gamma9 , and to extract multiplexing gains from the statistical independence of traffic processes. ... |

419 |
Equivalent capacity and its application to bandwidth allocation in high-speed networks
- Guerin, Ahmadi, et al.
- 1991
(Show Context)
Citation Context ... qualitative results for the purposes of real-time administration of admission control. A key concept here is effective bandwidth, which has recently received considerable attention for both buffered =-=[7, 8, 6, 21]-=- and unbuffered [10, 11, 14] network resources. Underlying the concept is linearity of @AL , the boundary of the admissible set AL , which is the set of combinations of sources of various classes for ... |

336 |
Sondhi: Stochastic Theory of a Data-Handling System with Multiple Sources
- Anick, Mitra, et al.
(Show Context)
Citation Context ...esources, a buffer of size B and a trunk of transmission bandwidth C. One approach to network capacity calculations and admission control has been based on statistical source models. See for instance =-=[1, 18]-=-, and also [5] for analysis of the combined effects of such models and regulators. The reliability of source models in this approach is a matter of concern, however. A contrasting approach is based on... |

268 | Effective bandwidth of general Markovian traffic sources and admission control of high speed networks
- Elwalid, Mitra
- 1993
(Show Context)
Citation Context ... qualitative results for the purposes of real-time administration of admission control. A key concept here is effective bandwidth, which has recently received considerable attention for both buffered =-=[7, 8, 6, 21]-=- and unbuffered [10, 11, 14] network resources. Underlying the concept is linearity of @AL , the boundary of the admissible set AL , which is the set of combinations of sources of various classes for ... |

219 |
New Directions in Communications (or Which Way to the Information Ap
- Turner
- 1986
(Show Context)
Citation Context ...ervice requirements combine to pose challenges of unprecedented difficulty. An essential prerequisite is the regulation of traffic at network edges, as shown in Figure 1.1. The Leaky Bucket regulator =-=[20]-=- has become ubiquitous in current service offerings. In our work we assume that all traffic is subject to regulation by such devices, which have three parameters: the token rate r, which bounds the lo... |

170 |
Resource Allocation for Broadband Networks
- Hui
- 1988
(Show Context)
Citation Context ... paper, where one of the two key network resources is the buffer, is novel in its use of established techniques for analyzing unbuffered resources, such as the Chernoff large deviations approximation =-=[9, 10, 11, 14]-=-. The important bridge to the case where bandwidth is the other network resource is based on a technique that allows the two resources to be exchangeable. The first phase of the analysis concludes wit... |

150 |
Switching and traffic theory for integrated broadband networks
- Hui
- 1990
(Show Context)
Citation Context ... paper, where one of the two key network resources is the buffer, is novel in its use of established techniques for analyzing unbuffered resources, such as the Chernoff large deviations approximation =-=[9, 10, 11, 14]-=-. The important bridge to the case where bandwidth is the other network resource is based on a technique that allows the two resources to be exchangeable. The first phase of the analysis concludes wit... |

96 |
Effective bandwidths for the multi-type UAS channel, Queueing Systems 9(1–2
- Gibbens, Hunt
- 1991
(Show Context)
Citation Context ... qualitative results for the purposes of real-time administration of admission control. A key concept here is effective bandwidth, which has recently received considerable attention for both buffered =-=[7, 8, 6, 21]-=- and unbuffered [10, 11, 14] network resources. Underlying the concept is linearity of @AL , the boundary of the admissible set AL , which is the set of combinations of sources of various classes for ... |

86 |
Tail probabilities with statistical multiplexing and e ective bandwidths in multi-class queues
- Whitt
- 1993
(Show Context)
Citation Context |

57 |
Analysis and design of rate-base congestion control of high speed networks, I:stochastic fluid models, access regulation”, Queueing Systems 9
- Elwalid, Mitra
- 1991
(Show Context)
Citation Context ... of size B and a trunk of transmission bandwidth C. One approach to network capacity calculations and admission control has been based on statistical source models. See for instance [1, 18], and also =-=[5]-=- for analysis of the combined effects of such models and regulators. The reliability of source models in this approach is a matter of concern, however. A contrasting approach is based on worst case co... |

51 |
Erlang Capacity and Uniform Approximations for Shared Unbuffered
- Mitra, Morrison
- 1994
(Show Context)
Citation Context ... paper, where one of the two key network resources is the buffer, is novel in its use of established techniques for analyzing unbuffered resources, such as the Chernoff large deviations approximation =-=[9, 10, 11, 14]-=-. The important bridge to the case where bandwidth is the other network resource is based on a technique that allows the two resources to be exchangeable. The first phase of the analysis concludes wit... |

41 |
Deterministic rule based traffic descriptors for broadband ISDN: Worst case behavior and connection acceptance control
- Doshi
- 1994
(Show Context)
Citation Context ...es. We do not have conclusive proof that such processes are worst-case in the sense of maximizing the steady-state loss probability in the ATM node. However, this is suggested by the results of Doshi =-=[4]-=-, Mitra and Morrison [15] and Worster [22] (see also [23]). Doshi has various related results in general frameworks. In [15] it is specifically shown that, for a single-resource problem, which is what... |

40 |
Performance Evaluation and Design of Multiservice Networks
- Roberts, editor
- 1992
(Show Context)
Citation Context ...esources, a buffer of size B and a trunk of transmission bandwidth C. One approach to network capacity calculations and admission control has been based on statistical source models. See for instance =-=[1, 18]-=-, and also [5] for analysis of the combined effects of such models and regulators. The reliability of source models in this approach is a matter of concern, however. A contrasting approach is based on... |

24 |
Strong large deviation and local limit theorems
- Chaganty, Sethuraman
- 1993
(Show Context)
Citation Context ...Pr(u jisx) : (3.5) Chernoff's bound gives [2], log P losss\GammaF K (s ) ; (3.6) where F K (s) = sC \Gamma J X j=1 K j log M j (s) (3.7) and F K (s ) = sup s0 F K (s). If C !1 and K j =C = O(1), then =-=[3], [1-=-7] log P loss = \GammaF K (s ) 1 + O ` log C C ' : (3.8) Hence the asymptotic large deviations approximation is P loss �� exp(\GammaF K (s )). To avoid trivialities assume the stability condition,... |

23 |
Traffic analysis in large-scale high-speed integrated networks:validation of nodal decomposition approach, INFOCOM
- Lau, Li
- 1993
(Show Context)
Citation Context ...y few other sources. However, when the source peak rate is small compared to the link rate, i.e., the number of sources multiplexed is large, the probability that only few sources are on is small. In =-=[13]-=- simulation studies using source models with exponentially-distributed on and off periods confirm that the the change in the characteristics of the source traffic due to buffering and multiplexing in ... |

19 |
On the probabilities of large deviations for sums of independent random variables
- Petrov
- 1965
(Show Context)
Citation Context ...jisx) : (3.5) Chernoff's bound gives [2], log P losss\GammaF K (s ) ; (3.6) where F K (s) = sC \Gamma J X j=1 K j log M j (s) (3.7) and F K (s ) = sup s0 F K (s). If C !1 and K j =C = O(1), then [3], =-=[17] log-=- P loss = \GammaF K (s ) 1 + O ` log C C ' : (3.8) Hence the asymptotic large deviations approximation is P loss �� exp(\GammaF K (s )). To avoid trivialities assume the stability condition, J X j... |

13 |
Effective Bandwidth at Multi-type Queues
- Kelly
- 1991
(Show Context)
Citation Context ...ity of AL , ~ A L; ~ K is a conservative bound to AL . Moreover, denoting ~ A L; ~ K = 8 ! : K : J X j=1 K j e jsC 9 = ; ; (3.17) the parameters e j which define ~ A L; ~ K are readily obtained [10], =-=[11]-=-: e j = log M j (~s ) ~ s + (log L)=C (j = 1; 2; : : : ; J) ; (3.18) where ~ s is where F ~ K (s) is maximized. The quantity e j may justifiably be called the effective bandwidth of class j traffic so... |

12 |
Bounds and approximations for the periodic on/off queue with applications to ATM traffic control
- Kvols, Blaabjerg
- 1992
(Show Context)
Citation Context ...nds [24], and lead to more efficient use of network resources. This is little prior statistical analysis of burst-scale congestion in the multiplexing of periodic, on-off source traffic. Examples are =-=[12]-=- and [19]. In these papers the analysis is approximate and the calculations of the overflow probability, as pointed out in [12], are quite laborious. Also, no attempt is made to carry the analysis thr... |

11 |
Performance limitations of leaky bucket algorithm for usage parameter control of bandwidth allocation methods
- Yamanaka, Sato, et al.
- 1992
(Show Context)
Citation Context ...are worst-case in the sense of maximizing the steady-state loss probability in the ATM node. However, this is suggested by the results of Doshi [4], Mitra and Morrison [15] and Worster [22] (see also =-=[23]-=-). Doshi has various related results in general frameworks. In [15] it is specifically shown that, for a single-resource problem, which is what we have in this paper after the initial phase of the ana... |

9 |
Modelling deterministic queues: The leaky bucket as an arrival process
- Worster
- 1994
(Show Context)
Citation Context ...such processes are worst-case in the sense of maximizing the steady-state loss probability in the ATM node. However, this is suggested by the results of Doshi [4], Mitra and Morrison [15] and Worster =-=[22]-=- (see also [23]). Doshi has various related results in general frameworks. In [15] it is specifically shown that, for a single-resource problem, which is what we have in this paper after the initial p... |

5 |
A traffic control framework for high speed data transmission
- Roberts, Bensaou, et al.
- 1993
(Show Context)
Citation Context ... and lead to more efficient use of network resources. This is little prior statistical analysis of burst-scale congestion in the multiplexing of periodic, on-off source traffic. Examples are [12] and =-=[19]-=-. In these papers the analysis is approximate and the calculations of the overflow probability, as pointed out in [12], are quite laborious. Also, no attempt is made to carry the analysis through to t... |

5 | On Per-Session End-to-End Delay and the Call Admission Problem for Real-Time Applications with QOS Requirements
- Yates, Kurose, et al.
- 1994
(Show Context)
Citation Context ...ource behavior with a quality of service guarantee in the form of a small loss probability, is taken in this paper. It should counter the overly conservative usage implied by the deterministic bounds =-=[24]-=-, and lead to more efficient use of network resources. This is little prior statistical analysis of burst-scale congestion in the multiplexing of periodic, on-off source traffic. Examples are [12] and... |

4 |
Independent regulated processes to a shared unbuffered resource which maximize the loss probability
- Mitra, Morrison
- 1994
(Show Context)
Citation Context ...usive proof that such processes are worst-case in the sense of maximizing the steady-state loss probability in the ATM node. However, this is suggested by the results of Doshi [4], Mitra and Morrison =-=[15]-=- and Worster [22] (see also [23]). Doshi has various related results in general frameworks. In [15] it is specifically shown that, for a single-resource problem, which is what we have in this paper af... |

1 | Bounds and approximations for the periodic odoff queue with applications to ATM traffic control - Kvols, Blaabjerg - 1992 |

1 | Commission of the European - Roberts - 1992 |

1 | Elwalid (S’89-M’90), for a photograph and biography please see page 1016 of this issue. Debasis Mitra (M’75-SM’82-F’89). for a photograph and biography please see page 936 of this issue. Robert H. Wentworth (M’93) received the B.S. and B A degrees in phys - Anwar |