Is-Haka Mkwawa

Multicast communication in grid computing networks with background traffic

Grid computing is a computational concept based on an infrastructure that integrates and collaborates the use of high end computers, networks, databases and scientific instruments owned and managed by several organisations. It involves large amounts of data and computing which require secure and reliable resource sharing across organisational domains. Despite its high computing performance orientation, communication delays between grid computing nodes is a big hurdle due to geographical separation in a realistic grid computing environment. Communication schemes such as broadcasting, multicasting and routing should, therefore, take communication delay into consideration. Such communication schemes in a grid computing environment pose a great challenge due to the arbitrary nature of its topology. In this context, a heuristic algorithm for multicast communication is proposed for grid computing networks with finite capacity and bursty background traffic. The scheme facilitates inter-node communication for grid computing networks and it is applicable to a single-port mode of message passing communication. The scheme utilises a queue-by-queue decomposition algorithm for arbitrary open queueing network models, based on the principle of maximum entropy, in conjunction with an information theoretic decomposition criterion and graph theoretic concepts. Evidence based on empirical studies indicates the suitability of the scheme for achieving an optimal multicast communication cost, subject to system decomposition constraints.

AN OPTIMAL NEIGHBOURHOOD BROADCASTING SCHEME FOR STAR INTERCONNECTION NETWORKS

A novel and efficient neighbourhood broadcasting scheme is proposed for star interconnection networks, based on binomial trees and graph theoretic concepts. The optimal scheme has an upper bound of 1.33⌈log2(n - 2)⌉ + O(1) time steps and it is applicable to both single-port and half-duplex modes of message passing communication.

Performance Modelling and Evaluation of Handover Mechanism in IP Multimedia Subsystems

IP multimedia subsystems (IMS) is using session initiation protocol (SIP) as signalling protocol. SIP is becoming computing intensive and hence the ability of entry and exit proxies to process SIP messages has great impact on the overall performance of the IMS core network. This paper shows that the flow of SIP messages during handover process in the IMS core network complies with the well known central server model. The later is validated via discrete event simulation results. The performance prediction study reveals that S-CSCF is the bottleneck device of the IMS core network processing a large number of SIP messages.

Broadcasting methods in MANETS: an overview

Broadcasting in mobile ad hoc networks (MANETs) is an information dissemination process of sending a message from a source node to all other nodes of the network. Even though it has been studied extensively for wired networks, broadcasting in MANETs poses more challenging problems because of the variable and unpredictable characteristics of its medium as well as the fluctuation of the signal strength and propagation with respect to time and environment. Furthermore, node mobility creates a continuously changing communication topology in which routing paths break and new ones form dynamically. In this context, efficient broadcasting in Mobile Ad hoc networks is crucial for providing control and routing information for multicast and point to point communication protocols. This paper presents an overview on the state of the art of broadcasting methods in MANETs and makes recommendations on how to improve the efficiency and performance of tree and cluster based broadcasting methods.

One-to-all broadcasting scheme for hypercubes with background traffic

A novel one-to-all broadcasting scheme is proposed for hypercube interconnection networks with bursty background traffic having head-of-line priorities, complete buffer sharing and repetitive service blocking with random destination. The scheme facilitates inter-process communication amongst hypercube computers and it is applicable to a single-port mode of message passing communication. The scheme utilizes a queue-by-queue decomposition algorithm for arbitrary open queueing network models, based on the principle of maximum entropy, in conjunction with binomial trees and graph theoretic concepts. Evidence based on empirical studies indicates the suitability of the scheme for achieving an optimal one-to-all broadcasting cost.

Performance modelling and evaluation of a mobility management mechanism in IMS-based networks

The 3rd Generation Partnership Project (3GPP) for an IP multimedia subsystem (IMS) architecture defined a number of functional units, which exchange session initiation protocol (SIP) messages with register users and set up or terminate multimedia sessions. The processing of SIP messages, however, requires a significant amount of processing, queueing and transmission times with adverse implications on the overall performance of the IMS core architecture. This tutorial introduces a mobility management mechanism for an IMS testbed implemented by Nokia-Siemens as part of the EU IST VITAL project. The mechanism employs an open queueing network model (QNM) with priorities to represent the functional units and application servers of an IMS architecture during a handover process of SIP messages between different access networks. The QNM is analysed via the principle of maximum entropy and numerical experiments are employed to assess the performance impact of bursty traffic flows of SIP messages on the IMS architecture.

Broadcasting schemes for hypercubes with background traffic

Optimal broadcasting schemes for interconnection networks (INs) are most essential for the efficient interprocess communication amongs: parallel computers. In this paper two novel broadcasting schemes are proposed for hypercube computers with bursty background traffic and a single-port mode of message passing communication. The schemes utilize a maximum entropy (ME) based queue-by-queue decomposition algorithm for arbitrary queueing network models (QNMs) [D.D. Kouvatsos, I. Awan, Perform. Eval. 51 (2003) 191] and are based on binomial trees and graph theoretic concepts. It is shown that the overall cost of the one-to-all broadcasting scheme is given by max{ω1, ω2.....ω2n/2}, where ωi i = 1, 2,..,2n/2 is the total weight at each leaf node of the binomial tree and n is the degree of the hypercube. Moreover, the upper bound of the total cost of the neighbourhood broadcasting scheme is determined by ΣFi = 1 max{ωi}, where F is an upper bound of the number of steps and is equal to 1.33[log2(n - 1)] + 1. Evidence based on empirical studies indicates the suitability of the schemes for achieving optimal broadcasting costs.

One-to-all broadcasting scheme for arbitrary static interconnection networks with bursty background traffic

A novel one‐to‐all broadcasting scheme is proposed for arbitrary static interconnection networks with bursty background traffic having head‐of‐line priorities, complete buffer sharing and repetitive service blocking with random destination. The scheme facilitates inter‐process communication amongst parallel computers and grid computing networks as well as asynchronous transfer mode space division architectures and it is applicable to a single‐port mode of message passing communication. The scheme utilizes a queue‐by‐queue decomposition algorithm for arbitrary open queueing network models, based on the principle of maximum entropy, in conjunction with an information theoretic decomposition criterion and graph theoretic concepts. Evidence based on empirical studies indicates the suitability of the scheme for achieving an optimal one‐to‐all broadcasting cost. Copyright

Overlay networks and graph theoretic concepts

Overlay networks have shown to be very effective towards the support and enhancement of network performance and the availability of new applications and protocols without interfering with the design of the underlying networks. One of the most challenging open issues in overlay networks, however, is paths overlapping, where overlay paths may share the same physical link and thus, the ability of overlay networks to quickly recover from congestion and path failures is severely affected. This chapter undertakes a review of some graph theoretic based methods for the selection of a set of topologically diverse routers towards the provision of independent paths for better availability, performance and reliability in overlay networks. Moreover, it proposes a graph decomposition-based approach for the maximization of path diversity without degrading network performance of in terms of latency. Some remarks on future developments and challenges in the field of overlay networks are included.

Multicast communication in interconnection networks with background traffic

The optimization of multicast communication for regular interconnection networks (IN) has been the subject of many studies in the literature. Extensions, however, the analysis of IN with arbitrary topology and background traffic poses a great challenge. In this context, a novel heuristic algorithm for multicast communication is proposed for arbitrary IN with finite capacity and bursty background traffic. The scheme facilitates interprocess communication amongst parallel computers and Grid computing networks as well as local and wide area networks and it is applicable to a single-port mode of message passing communication. The scheme utilizes a queue-by-queue decomposition algorithm for arbitrary open queueing network models, based on the principle of maximum entropy, in conjunction with an information theoretic decomposition criterion and graph theoretic concepts. Evidence based on empirical studies indicates the suitability of the scheme for achieving an optimal multicast communication cost.