Dynamic bandwidth allocation for ATM switches

Hsu, Ivy; Walrand, Jean

doi:10.2307/3215357

Cited by 30 publications

(20 citation statements)

References 7 publications

(7 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the short-range dependent context, several authors have proposed estimating θ * from observed traffic using the maximum workload r.v. M t ; see Berger and Whitt (1995) and Hsu and Walrand (1996). Theorem 1 states that if H is known, then θ * can also be successfully estimated from long-range dependent traffic using M t .…”

Section: Resultsmentioning

confidence: 99%

On the maximum workload of a queue fed by fractional Brownian motion

Zeevi¹

2000

Ann. Appl. Probab.

View full text Add to dashboard Cite

Consider a queue with a stochastic fluid input process modeled as fractional Brownian motion (fBM). When the queue is stable, we prove that the maximum of the workload process observed over an interval of length t grows like γ log t 1/ 2−2H , where H > 1/2 is the self-similarity index (also known as the Hurst parameter) that characterizes the fBM and can be explicitly computed. Consequently, we also have that the typical time required to reach a level b grows like exp b 2 1−H . We also discuss the implication of these results for statistical estimation of the tail probabilities associated with the steady-state workload distribution.1. Introduction. Triggered by measurements and statistical analysis of traffic in high-speed networks, recent research has focused on stochastic models of network traffic that have the properties of long-range dependence and self-similarity; see, for example, Leland, Taqqu, Willinger and Willson (1993), Beran, Sherman, Taqqu and Willinger (1995) and Erramilli, Narayan and Willinger (1997) for a discussion of the statistical evidence that favors such models. Perhaps the most theoretically important traffic model that exhibits these properties is fractional Brownian motion (fBM). In fact, just as Brownian motion is supported as a model of short-range dependency on the basis of Donsker's theorem and its generalizations, fBM can be viewed as a natural limiting approximation to a broad class of more physically plausible models that describe how traffic in a network is generated from its individual sources; see Samorodnitsky (1997, 1998), Konstantopoulos and Lin (1996) and Whitt (1998). Because of both theoretical and statistical evidence that supports fBM as a possible traffic model, there is significant interest in trying to reach an understanding of the implications of such long-range dependent traffic for the performance of queues (because queueing effects occur at the buffers to the switches in a network).Because fBM is highly non-Markovian (i.e., there is no finite-dimensional supplementary variable representation of fBM that makes it Markov), it is a challenging process to analyze as an input to a queue. At this point in time, we are aware of only two sets of results that describe the performance of a queue

show abstract

Section: Resultsmentioning

confidence: 99%

On the maximum workload of a queue fed by fractional Brownian motion

Zeevi¹

2000

Ann. Appl. Probab.

View full text Add to dashboard Cite

show abstract

“…Similar measurement-based schemes are considered in [10] and [15] to allocate bandwidth dynamically according to observed CLR at regular intervals. A virtual buffer (i.e.…”

Section: Measurement-based Approachmentioning

confidence: 99%

“…Here, the VOB uses a parallel set of counters to simulate virtual cells being multiplexed into several virtual buffers with different service capacities, as shown in Figure 1. It is conceptually similar to the parallel virtual buffers described in [10]. The major difference is that the target CLR in [10] must be large enough to be measured directly (e.g., 10 -3 ).…”

Section: Virtual Output Buffer Approachmentioning

confidence: 99%

“…It is conceptually similar to the parallel virtual buffers described in [10]. The major difference is that the target CLR in [10] must be large enough to be measured directly (e.g., 10 -3 ). On the contrary, we do not have this limitation in our case because the small target CLR (e.g., 10…”

Section: Virtual Output Buffer Approachmentioning

confidence: 99%

“…In [10][15], they have studied dynamic bandwidth allocation based on measured cell loss ratios (CLR). The measurement time interval of these methods is inversely proportional to the CLR.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Dynamic bandwidth allocation for real-time VBR video traffic in ATM networks

Chu

Tsang²

Proceedings of Sixth International Conference on Computer Communications and Networks

View full text Add to dashboard Cite

show abstract

Intelligent Flow Control under a Game Theoretic Framework

Chen

2001

Telecommunications Optimization: Heuristic and Adaptive Techniques

View full text Add to dashboard Cite

Each of us is interested in optimization, and telecommunications. Via several meetings, conferences, chats, and other opportunities, we have discovered these joint interests and decided to put together this book.It certainly wasn't easy. What made it difficult is partly the reason for this book in the first place: things are moving fast! As researchers in both industry and academia, we are under constant pressure to keep up with new technologies and opportunities. Part of our day jobs involves developing good algorithms for a steady surge of new problems arising from industry. Another part of our day jobs involves finding realistic problems which might yield to some of the algorithms we have developed, which we particularly like, but which don't work on anything else. The remainder of our day jobs tends to involve lecturing (for two of us), project management (for all of us), and things like exam-setting and marking. Finally, for two of us, childcare takes up a substantial part of our night jobs. Naturally, the book was finished when we worked out how to stuff 46 hours into a single day, and keep that up for a few successive weeks.Seriously, though, telecommunicatons is growing and changing very quickly, and all of this involves the increasing availability of new real-world problems for optimization researchers to address. Telecommunications engineers tend to know all about the problems involved, but are often not aware of developments in computer science and artificial intelligence which might go a long way towards solving those problems. We therefore argue, at the beginning of Chapter 1, that there needs to be more collaboration between the communities, so that both sides can be equipped for the challenges ahead. Indeed, this book is the result of one such collaboration.We have oriented it slightly towards the telecommunications engineer who knows little about the computational techniques, rather than the other way around. This is mainly because it is much easier to describe the techniques than it is to describe the telecommunications issues; the latter is a much larger collection. However, individual chapters go into some detail, often at an introductory level, in describing particular kinds of problem. Also, we provide a brief introduction to the essence of the problems at the beginning of each part.Let us know what you think. In particular, if you are a telecommunications engineer, and have been inspired by a chapter or two in this book to try out some techniques that you otherwise wouldn't have, please get in touch: your story may make it into the next edition!

show abstract

Dynamic bandwidth allocation for ATM switches

Cited by 30 publications

References 7 publications

On the maximum workload of a queue fed by fractional Brownian motion

On the maximum workload of a queue fed by fractional Brownian motion

Dynamic bandwidth allocation for real-time VBR video traffic in ATM networks

Intelligent Flow Control under a Game Theoretic Framework

Contact Info

Product

Resources

About