Network Optimization Using Evolutionary Algorithms in Multicast Transmission

Network Optimization Using Evolutionary Algorithms in Multicast Transmission

Yezid Donoso (Universidad del Norte, Colombia) and Ramón Fabregat (Girona University, Spain)
Copyright: © 2008 |Pages: 7
DOI: 10.4018/978-1-59140-993-9.ch048
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

To support QoS in today’s Internet, several new architecture models have been proposed (Striegel, A., & Manimaran, G. (2002)). Traffic engineering has become a key issue within these new architectures, as supporting QoS requires more sophisticated resource management tools. Traffic engineering aims to optimize the performance of operational networks. The main objective is to reduce congestion hot spots and improve resource utilization. This can be achieved by setting up explicit routes through the physical network in such a way that the traffic distribution is balanced across several traffic trunks. This load balancing technique can be achieved by multicommodity network flow (Pioro, M., & Medhi, D. (2004)) formulation. This leads to the traffic being shared over multiple routes between the ingress node and the egress nodes in order to avoid link saturation and hence the possibility of congestion, which is the inability to transmit a volume of information with the established capacities for a particular equipment or network.

Key Terms in this Chapter

Minimize Maximum Link Utilization: When the maximum link utilization is minimized, represented by a in some cases, a new upper bound of the utilization in every link of the network is created. In this way, when this new upper bound is exceeded, the information flow is transmitted by another different path. That is, for all the traffic that exceeds the (a.uij), where uij is the link capacity, the value will be transmitted by other paths instead of using the total capacity of the links. If the traffic is multicast, instead of paths we consider trees.

Load Balancing: A fundamental mechanism for implementing traffic engineering. It concerns the number of overutilized links, which will have low QoS, and underutilized links, which represent a waste to be reduced. It can be achieved by a multicommodity network flow formulation which leads to the traffic being shared over multiple routes between the ingress node and the egress nodes in order to avoid link saturation and hence the possibility of congestion.

Multicast Connections: Transmission from one source to a set of destinations. The aim of multicasting is to be able to send data form a sender to the set of destinations in an efficient manner.

Multipath Routing: Different paths can be used to provide a balanced traffic over the links. Links do not get overused and so do not get congested. Therefore, they have the potential to aggregate bandwidth, allowing a network to support a higher data transfer than is impossible with any single path.

Evolutionary Algorithm (EA): Evolutionary algorithms (EA) are inspired by Darwin’s theory of evolution, which is based on the survival of the fittest species. The character of a life is determined by the chromosomes. In a chromosome there are many different genes, which indicate different characters. The EA principle is based on the different combinations of genes in a chromosome. Different combinations will lead to different characters and the more suitable ones will remain in the world.

Quality of Service (QoS): QoS is defined as a collective effect of service performance, which determines the degree of satisfaction of a user of the service. It is the capability of a network to provide better services to selected network traffic over various (heterogeneous) technologies. QoS is a set of service requirements to be met by the network while transporting a flow.

Traffic Engineering (TE): Concerned with performance optimization of operational networks. In general, it encompasses applying technology and scientific principles to measuring, modelling, characterizing, and controlling Internet traffic, and applying the knowledge and techniques to achieve specific performance objectives.

Multiprotocol Label Switching (MPLS): MPLS integrates the label switching forwarding paradigm with network layer routing. MPLS is an approach to packet forwarding whereby short fixed length labels are attached to IP packets by a LSR at the ingress to an MPLS domain. An ingress LSR attaches labels to packets based on the concept of forwarding equivalence classes (FEC), so that packets that belong to the same FEC are assigned the same label value and follow the same LSP.

Label Switched Path (LSP): It is a path between source node and the egress node in MPLS technology.

Multiobjective Problem (MOP): A general MOP includes a set of n parameters (decision variables), a set of k objective functions, and a set of m restrictions. The objective and restriction functions are functions of the decision variable. In multiobjective problems, it rarely happens that all the objectives can be optimized simultaneously; it is generally the case that the objectives conflict with each other. Because the multi-objective optimization problem does not always have a single solution and the problem is known to be NP-complete, in several works EA solutions are proposed.

Memetic Algorithm: As the evolutionary algorithms are algorithms that can give a solution to combinatorial problems of the same way than the natural behaviour.

Complete Chapter List

Search this Book:
Reset