Parsimonious estimates of bandwidth requirement for quality of service packet networks

Proceedings of the 1st International Conference on Performance Evaluation Methodolgies and Tools - Valuetools '06

2006

The estimation of the expected traffic loss ratio (workload loss ratio, WLR) is a key issue in provisioning Quality of Service in packet based communication networks. The stationary (long run) loss ratio in queuing analysis is usually estimated through other assessable quantities, typically based on the approximates of the buffer overflow probability. These approaches have the drawback, that the relation between loss ratio and buffer overflow probability is often hardly quantifiable and it can in principle be arbitrary. In this paper we present novel upper approximations for workload loss ratio derived from the original definition of stationary loss ratio. These direct bounds are applicable for general service curve network elements with regulated flows acting as inputs. The performance of these bounds is systematically analyzed and compared to prior indirect (based on buffer saturation probability) loss ratio bound and its newly developed improvements. The extensive numerical investigations also exemplify that in most cases our novel direct bounds can lead to significant save in buffer requirements when guaranteeing a prescribed QoS level in terms of loss ratio.

Section: Wlr Estimation With Direct Formulaementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Loss ratio approximations in buffered systems with regulated inputs

Proceedings of the 1st International Conference on Performance Evaluation Methodolgies and Tools - Valuetools '06

2006

“…From (7) it can be seen, that giving a better bound on the PGF, gives a better bound on the backlog as well. In the proof of Theorem 1 of [7], there is a step which states, that if:…”

Section: Approximation Of the Probability Generating Function (Pgfmentioning

confidence: 99%

“…To create the new bounds on the backlog, we will need the result of the following theorem presented in [7], and [2]: …”

Section: This Is Becausementioning

confidence: 99%

Buffer overflow estimation in network elements, multiplexing independent regulated inputs

Gulyás

Szenasi

IEEE Global Telecommunications Conference, 2004. GLOBECOM '04.

The estimation of resource usage in network nodes is a basic issue from the point of QoS (Quality of Service) provisioning. This paper is about estimating the buffer overflow within network elements, which are represented by a service curve property. There are some previous investigations in [1], which define bounds on buffer overflow for such elements, multiplexing homogeneous and heterogeneous independent regulated inputs. The main topic of this paper is setting up new bounds for the heterogeneous case, by using a new PGF (Probability Generating Function) approximation presented in [7]. We show, that the new bounds are also valid for the homogeneous case, so the results for the homogeneous and the heterogeneous case can be integrated into a single formula. It will also be shown, that the new bounds improve the existing ones in most cases.

A Novel Direct Upper Approximation for Workload Loss Ratio in General Buffered Systems

NETWORKING 2005. Networking Technologies, Services, and Protocols; Performance of Computer and Communication Networks; Mobile A

Gulyás

Heszberger

2005

Self Cite

The workload loss ratio (WLR) is a key quantity from the point of Quality of Service (QoS) provisioning in packet-based communication, hence it's estimation is an important issue. The existing results in the area of WLR approximation usually interpret the workload loss as a product of some well assessable quantities. We call this approach as the indirect approximation of the WLR. The drawback of this approach is that each estimation has an error and the product of these errors could result in a highly inaccurate bound. This work deals with the upper approximation of the workload loss ratio based on it's original definition and proposes a new direct bound on the WLR applicable in general service curve network element. Extensive and systematic performance analysis of the formulae have been done, in which we found that the direct approach gives more accurate bound in several cases. We illustrate this analysis through some numerical examples.