Network Services

Network Services

DOI: 10.4018/978-1-4666-7312-0.ch013
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

A novel framework formed from a collection of independent agents that interact with each other is determined to provide a network service. Agents in this structure have the capability to perform independent activities such as duplication, migration, etc. A new method is developed in this chapter by means of genetic algorithms to change the behavior of agents over peers and also to improve the network service performance in a distributed and well planned way. Architecture with a remote control device, Personal Universal Controller (PUC), is described. The PUC provides two-way communication with the applications for copying specification for its functionality and constructing an interface for monitoring that electrical device. The requirements of every application hold the information about its dependency information and availability of appliance conditions. The network protocols, such as Service Discovery Protocols, are explained with their types and functionality.
Chapter Preview
Top

Background

Adaptive Network Services

Recently, the fields like optimization, clustering, communication networks, robotics, etc., utilize the concept of swarm intelligence immensely. Swarm intelligence is intelligence which is observed in the group of simple insects like ants, birds etc. According to this, the information about the environment is only partially used instead the operation is carried out the individuals based on their capability of sensing their environment. The objective can be satisfied only by the intellectual behavior of the group of folks but the individuals will not be able to achieve the goal. The group of individuals offers the network services like adaptation and scalability.

The services of distributed system which comprise contented distribution networks, content service networks, distributed record offering system, etc., and need a bulk of network components to be duplicated, migrate, and remove in a decentralized procedure (Nakano, & Suda, 2004).

In this framework, the group of agents offers only one network service. Even though, every agent individually employs the same network service, they exhibit various behaviors. The behavior of the agent depends on the set of genes which are induced into it and the agent from which it is developed through reproduction process in genetic algorithms. There exists a central control in generational genetic algorithms which selects and evaluates the individual agents whereas in the evolutionary framework, the neighbor agents are involved in the evaluation and selection in a decentralized and autonomous mode. The behavior of the agents is advanced over generations and adapt to different network environments in evolutionary framework.

Evolutionary Framework for Developing Network Applications

The control in the network environment is distributed and the network service is controlled individually by each agent. Here, only the local agents can communicate with each other. Local agents are the agents that belong to the same network platform or the platforms which are adjacent to each other.

The three network components used in the modeling of the network are agents, users and platforms are shown in the Figure 1 as in (Nakano, & Suda, 2004). The energy is the common resource which exchanged among these components. The users get the network service from the agents and give energy to the agents. The platform where the agents reside is referred as the hosting platform. The hosting platform offer computing resources like CPU power, memory and network bandwidth to agents in order to exchange the energy. The actions like reproduction, migration and deletion are performed by the agents using this energy. The efficiency with which the service is offered by the agent using the computing resources and with which the action is performed can be measured by the level of energy. The independent systems connected to each other are referred as platforms. Agents reside on these platforms and the computing resources like CPU, memory, bandwidth, etc., are provided by the platforms. The agents perform the actions like reproduction, migration and deletion using the energy provided by the users on the hosting platform. The agents with no energy are ejected by the platforms. In this way, the energy efficient agents are supported and are selected by the platforms. The exchange of energy efficiently can be observed in Figure 1.

Figure 1.

Exchange of energy

Complete Chapter List

Search this Book:
Reset