Efficient bounds for bufferless statistical multiplexing

Heszberger, Zalán; Zatonyi, J.; Bı́ró, József; Henk, T.

doi:10.1109/glocom.2000.892095

Cited by 5 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Taking a closer look at the results, however, it can be seen, that apart from (5), the formulae are in implicit form or contain optimization tasks to be performed, which is a significant obstacle in using them in real-time environment. Although for the overflow probability there exist a few solution to overcome the difficulties in the evaluation [13], the attempt to yield upper bounds for C equ in closed form, or even to make appropriate approximations to get one have met several difficulties. In these cases, to reach acceptable results, numerical approximation techniques are inevitable to use, yielding sometimes prohibitively increased processing time.…”

Section: Computational Issues Of P Sat and C Equmentioning

confidence: 98%

Fast algorithms for computing parsimonious estimates of QoS measures

Heszberger

Bı́ró

Németh

2007

2007 International Symposium on Communications and Information Technologies

View full text Add to dashboard Cite

In recent publications concerning quality of service (QoS) measures in communication networks several attempts have been made to improve the techniques of computing or estimating the quality of data transfer. Many of the promising results are based on the knowledge of only very few parameters on the traffic situation, some of which are known a priory, others are measured. The paper deals with such parsimonious estimation techniques of QoS measures under the bufferless fluid flow multiplexing (bffm) framework, and concentrates on the efficient computation of them. The proposed fixed-point recursive algorithms are actively used to put the implicit formulae or ones containing optimization tasks into real-time environment.

show abstract

Section: Computational Issues Of P Sat and C Equmentioning

confidence: 98%

Fast algorithms for computing parsimonious estimates of QoS measures

Heszberger

Bı́ró

Németh

2007

2007 International Symposium on Communications and Information Technologies

View full text Add to dashboard Cite

show abstract

“…Taking a closer look at the results, however, it can be seen, that apart from ( 5), the formulae are in implicit form or contain optimization tasks to be performed, which is a significant obstacle in using them in real-time environment. Although for the overflow probability there exists a few solution to overcome the difficulties in the evaluation [14], the attempts to yield upper bounds for C equ in closed form, or even to make appropriate approximations of them encounter several difficulties. In these cases, to reach acceptable results, numerical approximation techniques are inevitable to use, yielding sometimes prohibitively increased processing time.…”

Section: Computational Issues Of P Sat and C Equmentioning

confidence: 99%

Fast Algorithms for Computing Parsimonious Estimates of QoS Measures

Bı́ró¹

2008

ECTI-EEC

View full text Add to dashboard Cite

In recent publications concerning quality of service (QoS) measures in communication networks several attempts have been made to improve the techniques of computing or estimating the quality of data transfer. Many of the promising results are based on the knowledge of only very few parameters on the traffic situation, some of which are known a priory, others are measured. The paper deals with such parsimonious estimation techniques of QoS measures under the bufferless fluid flow multiplexing (bffm) framework, and concentrates on the efficient computation of them. The proposed fixed-point recursive algorithms are key elements in applying either the implicit formulae or ones containing optimization tasks in real-time environment.

show abstract

“…Also, we note paper [9] where similar exponential bounds were constructed for Markov chains. Generally, exponential bounds proved to be very effective in order to define required size of the sample in order to ensure proper quality of estimation, see, for example, [11,7,6,15,17].…”

Section: Introductionmentioning

confidence: 99%

On the Exponential Probability Bounds for the Bernoulli Random Variables

Nikulin

2010

Preprint

View full text Add to dashboard Cite

We consider upper exponential bounds for the probability of the event that an absolute deviation of sample mean from mathematical expectation p is bigger comparing with some ordered level ε. These bounds include 2 coefficients {α, β}. In order to optimize the bound we are interested to minimize linear coefficient α and to maximize exponential coefficient β. Generally, the value of linear coefficient α may not be smaller than one. The following 2 settings were proved: 1) {1, 2} in the case of classical discreet problem as it was formulated by Bernoulli in the 17th century, and 2) {1, 2 1+ε 2 } in the general discreet case with arbitrary rational p and ε. The second setting represents a new structure of the exponential bound which may be extended to continuous case.

show abstract

Efficient bounds for bufferless statistical multiplexing

Cited by 5 publications

References 11 publications

Fast algorithms for computing parsimonious estimates of QoS measures

Fast algorithms for computing parsimonious estimates of QoS measures

Fast Algorithms for Computing Parsimonious Estimates of QoS Measures

On the Exponential Probability Bounds for the Bernoulli Random Variables

Contact Info

Product

Resources

About