Alan Grigg

Controlling mean queuing delay under multi-class bursty and correlated traffic

This paper presents an adaptive queue management scheme to maintain queuing delay in a router at a required level based on a comprehensive analytical model under aggregated Internet traffic flows from various traffic classes. The proposed scheme uses a closed-loop feedback control mechanism to constrain the average queuing delay by regulating traffic arrival rate implicitly through a movable queuing threshold. A discrete-time queuing model is developed to derive the relationship between average queuing delays and queuing thresholds based on a traffic model that models aggregated Internet traffic through superposition of N MMBP-2 arrival processes. The queuing threshold is adjusted dynamically with reference to the relationship derived in the analytical model and also feedback of average queuing delay measurement. Packets are dropped dynamically with respect to the changes of queuing threshold and the packet loss events serve as implicit congestion indicators. Matlab is used to perform queuing analysis and simulation. Statistical evaluation is performed to show the efficiency and accuracy of the analytical and simulation results.

QoS Enhancements and Performance Analysis for Delay Sensitive Applications

This paper presents a comprehensive system modeling and analysis approach for both predicting and controlling queuing delay at an expected value under multi-class traffic in a single buffer. This approach could effectively enhance QoS delivery for delay sensitive applications. Six major contributions are given in the paper: (1) a discrete-time analytical model is developed for capturing multi-class traffic with binomial distribution; (2) a control strategy with dynamic queue thresholds is used in simulation experiments to control the delay at a specified value within the buffer; (3) the feasibility of the system is validated by comparing theoretical analysis with simulation scenarios; (4) the arrival rate can be adjusted for each forthcoming time window during the simulation with multi-packet sources; (5) statistical evaluation is performed to show both efficiency and accuracy of the analytical and simulation results; (6) a graphical user interface is developed that can provide flexible configuration for the simulation and validate input values.

ADDER: Probabilistic, Application Layer Service Discovery for MANETs and Hybrid Wired-Wireless Networks

Over the past years, Mobile ad-hoc networks (MANETs) have attracted a considerable degree of research attention, with service discovery, selection and invocation being among the topics of interest of previous efforts. In this paper we introduce ADDER, a probabilistic, hybrid, directory-less service discovery mechanism. It has been designed for military IPv6-based MANETs but will work in any hybrid wired-wireless deployment. It achieves very low service acquisition time through the exchange of a very small number of short messages. Propagation of service descriptions is based on a distance vector algorithm, achieving loop and starvation freedom through a feasibility condition, which has been adopted from established and well-tested routing protocols. This paper also presents evaluation results, obtained by actual execution of the ADDER daemon on two different test beds. The experiments aim to demonstrate that the mechanism achieves good scalability with increasing number of services and network size.

RED and WRED Performance Analysis Based on Superposition of N MMBP Arrival Proccess

Random Early Detection (RED) is recommended as the congestion avoidance mechanism in packet switched networks in RFC2309 and has become the focus of much research into congestion control. Furthermore, RED and WRED are the AQM schemes most widely deployed in commercial Internet routers, such as Cisco and Juniper routers. This paper describes an analytical approach that models RED and WRED schemes using a discrete-time queuing model with superposition of N MMBP-2 sources as arrival processes. Each MMBP-2 represents an aggregated Internet traffic flow from different traffic classes or with different precedence. Performance analysis for RED and WRED are then carried out using different dropping profiles through Matlab numerical analysis based on the proposed queuing model. Performance measures, such as throughput, mean queuing delay and loss probability, obtained from analytical analysis are studied; the trade-off between packet loss and queuing delay can then be evaluated easily for different traffic classes. In this way, the proposed analytical model facilitates fine tuning of RED and WRED parameters for the purpose of QoS provisioning and capacity planning on Internet routers.

Delay Restraining of Combined Multiple Input Cross Core Router

Delay is arguably one of the most vital Quality of Service (QoS) metrics in computer networks since it has many implications for real time multimedia applications. The congestion and queuing delay of Internet core routers is one of the most important fundamental research topics which plays a vital role in determination of the overall end-to-end delay. In previous works, the control algorithm proposed and modelled in the discrete-time domain, by controlling the value of the threshold within the queue buffer to achieve the required delay, is based on a single arrival in one time slot. In this paper, a new traffic model with the ability to capture the multiple arrivals in a single time slot is described and its application for restraining the delay to a pre-defined value within the network core router is developed. The feasibility of the system is examined using both theoretical analysis and simulation. A statistical evaluation is performed to show both efficiency and accuracy of the results acquired throughout the simulation.

Bounding Queuing Delay in a Router Based on Superposition of N MMBP Arrival Process

This paper presents a comprehensive analytical approach to maintain queuing delay in a router at a required level through closed-loop feedback control under aggregated traffic flows from various traffic classes. The approach is able to bound average queuing delay but still achieve a high system utilization by regulating traffic arrival rate implicitly through a movable queuing threshold. The queuing threshold is adjusted dynamically through feedback of average queuing delay mea- surement. Consequently, packets are dropped dynamically with respect to the changes of queuing threshold and the packets loss event served as an implicit congestion indicator. Three major contributions are given in this paper: (1) a discrete time queuing model is developed to derive the mapping table of average queuing delay and queuing threshold; (2) a traffic model that models LRD aggregated Internet traffic through superposing of N MMBP-2 arrival processes; (3) a closed-loop feedback control to control queuing delay at its target level based on expected system utilization. Matlab is used to perform queuing analysis and simulation. Statistical evaluation is performed to show the efficiency and accuracy of the analytical and simulation results. I. INTRODUCTION

Performance analysis of a threshold-based discrete-time queue using maximum entropy

The provision of guaranteed QoS for various Internet traffic types has become a challenging problem for researchers. New Internet applications, mostly multimedia-based, require differentiated treatments under certain QoS constraints. Due to a rapid increase in these new services, Internet routers are facing serious traffic congestion problems. This paper presents an approximate analytical performance model in a discrete-time queue, based on closed form expressions using queue threshold, to control the congestion caused by the bursty Internet traffic. The methodology of maximum entropy (ME) has been used to characterize closed form expressions for the state and blocking probabilities. A discrete-time GGeo/GGeo/1/{N1, N2} censored queue with finite capacity, N2, external compound Bernoulli traffic process and generalised geometric transmission times under a first come first serve (FCFS) rule and arrival first (AF) buffer management policy has been used for the solution process. To satisfy the low delay along with high throughput, a threshold, N1, has been incorporated to slow the arrival process from mean arrival rate k1 to k2 once the instantaneous queue length has been reached, otherwise the source operates normally. This creates an implicit feedback from the queue to the arrival process. The system can be potentially used as a model for congestion control based on the Random Early Detection (RED) mechanism. Typical numerical experiments have been included to show the credibility of ME solution against simulation for various performance measures and to demonstrate the performance evaluation of the proposed analytical model.

A quality of service framework for dependability in large-scale distributed systems

As recognition grows within industry for the advantages that can be gained through the exploitation of large-scale dynamic systems, a need emerges for dependable performance. Future systems are being developed with a requirement to support mission critical and safety critical applications. These levels of criticality require predictable performance and as such have traditionally not been associated with adaptive systems. The software architecture proposed for such systems takes its properties from the service-oriented computing paradigm and the communication model follows a publish/subscribe approach. While adaptive, such architectures do not, however, typically support real-time levels of performance. There is scope, however, for dependability within such architectures through the use of Quality of Service (QoS) methods. QoS is used in systems where the distribution of resources cannot be decided at design time. In this paper a QoS based framework is proposed for providing adaptive and dependable behaviour for future large-scale dynamic systems through the flexible allocation of resources. Simulation results are presented to demonstrate the benefits of the QoS framework and the tradeoffs that occur between negotiation algorithms of varying complexities.

Review: Cross-layer signalling and middleware: A survey for inelastic soft real-time applications in MANETs

This paper provides a review of the different cross-layer design and protocol tuning approaches that may be used to meet a growing need to support inelastic soft real-time streams in MANETs. These streams are characterised by critical timing and throughput requirements and low packet loss tolerance levels. Many cross-layer approaches exist either for provision of QoS to soft real-time streams in static wireless networks or to improve the performance of real and non-real-time transmissions in MANETs. The common ground and lessons learned from these approaches, with a view to the potential provision of much needed support to real-time applications in MANETs, is therefore discussed.

A quality of service framework for adaptive and dependable large scale system-of-systems

There is growing recognition within industry that for system growth to be sustainable, the way in which existing assets are used must be improved. Future systems are being developed with a desire for dynamic behaviour and a requirement for dependability at mission critical and safety critical levels. These levels of criticality require predictable performance and as such have traditionally not been associated with adaptive systems. The software architecture proposed for such systems is based around a publish/subscribe model, an approach that, while adaptive, does not typically support critical levels of performance. There is, however, the scope for dependability within such architectures through the use of Quality of Service (QoS) methods. QoS is used in systems where the distribution of resources cannot be decided at design time. A QoS based framework is proposed for providing adaptive and dependable behaviour for future large-scale system-of-systems. Initial simulation results are presented to demonstrate the benefits of QoS.