Mobility-Aware Grid Computing

Mobility-Aware Grid Computing

Konstantinos Katsaros (Athens University of Economics and Business, Greece) and George C. Polyzos (Athens University of Economics and Business, Greece)
DOI: 10.4018/978-1-60566-026-4.ch419
OnDemand PDF Download:
$37.50

Abstract

Grid computing has emerged as a paradigm for coordinated resource sharing and problem solving in dynamic, multiinstitutional virtual organizations (Foster, 2001). A grid computing system is essentially a large-scale distributed system designed to aggregate resources from multiple sites, giving to users the opportunity to take advantage of enormous computational, storage, or bandwidth resources that would otherwise be impossible to attain. Current applications of grid computing focus on computational-expensive processing of large volumes of scientific data, for example, for earthquake simulation, signal processing, cancer research, and pattern search in DNA sequences. At the same time, the recent advances in mobile and wireless communications have resulted in the availability of an enormous number of mobile computing devices such as laptop PCs and PDAs (personal digital assistants). Thus, it is natural to extend the idea of resource sharing to mobile and wireless computing environments. Resource-sharing collaboration between mobile users appears as a promising research direction toward the alleviation of the inherent resource constraints present in mobile computing environments. Either in the context of mobile ad hoc networks (MANETs) or in wireless networks based on fixed infrastructure (i.e., cellular networks, wireless local area networks (WLANs), small- or large-scale communities of mobile users can form mobile grid systems and collaborate in order to either achieve a common goal (otherwise impossible to achieve) or simply overcome their individual limitations. In the following, we highlight the fundamental issues toward the realization of a computational mobile grid system.
Chapter Preview
Top

Background

The research area of mobile grid is relatively new compared to the traditional (fixed) grid and mobile and wireless computing research areas. A consensus on the exact character of mobile grid computing has not been reached yet. Hence, a classification of the existing approaches is provided in the following, aiming at the clarification of the mobile grid concept. At the same time, the most significant research directions within each approach are described.

As mentioned earlier, a primary distinction is made between various research efforts in the area based on whether mobile devices act exclusively as resource consumers or as resource providers as well.

Mobile Devices as Resource Consumers

In this case, research is motivated by the fact that mobile devices are considered to have limited computational and/or storage capabilities (Banavar, Beck, Gluzberg, Munson, Sussman, & Zukowski, 2000; Migliardi, Maheswaran, Maniymaran, Card, & Azzedin, 2002; Park, Ko, & Kim, 2003; Srinivasan, 2005). The grid, in this case, can provide the resources missing in mobile devices on demand. The emerging problems here stem from the mobile and wireless character of the devices and include intermittent connectivity, device heterogeneity (in terms of hardware and operating system), and limited battery life. The use of proxies is proposed in Park et al., which act as gateways to the grid. These proxies undertake the role of the mediator between the mobile device and the grid system, and try to hide device heterogeneity and intermittent connectivity by acting on behalf of the mobile device.

Other approaches target the provision of a “smart” environment for pervasive computing (Banavar et al., 2000; Srinivasan, 2005). Here, mobile devices are considered as pure access devices with no need for enhanced processing and/or storage capabilities (Migliardi et al., 2002). The role of the grid is to provide all the functionalities required by users, pushing this way the complexity of the system to the network rather than to the edges.

Key Terms in this Chapter

Reputation Mechanism: It is a mechanism used in P2P systems to associate the identities of peers with the opinions of other peers about the contribution of the former, and to acquaint the participating peers with this association. It is mostly used to discourage free riding.

Replication: Replication involves introducing redundant replicas of a resource in a system in order to improve reliability, performance, availability, and/or fault tolerance.

Mobile Grid Computing: It is the computing paradigm for coordinated resource aggregation and sharing in which mobile computing devices act either as resource consumers or as resource providers or both.

Accounting Mechanism: This is the process of recording (a summary of) the details of service consumption that usually follows successful authentication and authorization.

Incentive Scheme: It is a mechanism used to provide the motive to the actors of a system for a certain behavior.

Load Balancing: Load balancing is the technique used to spread workload or resource consumption between many resource providers in order to improve resource utilization and/or performance.

Free Riding: Free riding is a problem in P2P (peer-to-peer) systems, in which peers tend not to contribute resources in order to minimize their own costs while at the same time benefiting from the contributions of other peers.

Complete Chapter List

Search this Book:
Reset