Two-Tier Fault-Tolerant Exactly-Once Multicast Protocol for Distributed Mobile Systems: Design and Evaluation
With the advent of distributed mobile systems, many new mobile commerce applications such as mobile auctions, will gain benefit if group communication among users is supported by the wired and wireless networks. Such applications require atomic all-or-none semantics, as well as secure and reliable multicast. The key constraints of the mobile host, namely poor computational resources, limited battery power, low bandwidth, and problems due to mobility, has made multicast communication in distributed mobile system a challenging goal. To support reliable multicast, the protocol could allow one or more different retransmissions schemes. Keeping the constraints in mind it is mandatory to see to that the mobile hosts ideally avoid the repeated reception of the data. This characteristic led us to investigate reliable multicast protocol that could deliver the message exactly-once for distributed mobile systems. The novel feature of the protocol is viewing the mobile systems model as a two-tier model (unlike the three-tier existing view), reducing the communication overhead in the wired network, especially by not taking any host as a centralized host. Thus it reduces the risk of a single point failure. The protocol tolerates mobile support station and mobile host failures, and provides total ordered message delivery
Moset: An Anonymous Remote Mobile Cluster Computing Paradigm
The advance of technology in terms of cellular communications and the increasing computing power of the mobile systems have made it convenient for people to use more of mobile systems rather than static systems. This has seen more of mobile devices in personal and distributed computing, thus making the computing power ubiquitous. The combination of wireless communication and cluster computing in many applications has led to the integration of these two technologies to emerge as Mobile Cluster Computing (MCC) paradigm. This has made parallel computing feasible on mobile clusters, by making use of the idle processing power of the static and mobile nodes that form the cluster. To realize such a system for parallel computing, various issues such as connectivity, architecture and operating system heterogeneities, timeliness issues, load fluctuations on machines, machine availability variations and failures in workstations and network connectivities need to be handled. Moset, an Anonymous Remote Mobile Cluster Computing (ARMCC) paradigm is being proposed to handle these issues. Moset provides transparency to mobility of nodes, distribution of computing resources and heterogeneity of wired and wireless networks. The model has been verified and validated by implementing a distributed image rendering algorithm over a simulated mobile cluster model.
An Object Based Paradigm for Integration of Mobile Hosts into Grid: A Service Composition Approach
This paper proposes a perspective of a novel object mobile grid as an integration of mobile computing paradigm into the grid computing paradigm along with service composition technology. It shares both advantages of powerful computation capability of grid and ubiquitous accessibility of distributed mobile system. It combines the middleware solution from services composition, resource-sharing solutions of grid computing and the anywhere, anytime resource access of mobile computing. It is a distributed system model where computational, data and other resources including the experimental devices throughout the wired and wireless networks are organized into federation, as peer to peer model. All the resources, data and services offered by the participating static and mobile hosts are virtualized as services, to enable service composition which may require both resources and services for composition of a complex service. The proposed system is realized as a shared distributed object space enabling the scalability, and helps in handling the constraints of the distributed mobile system. The benefits of this model are information processing capacity increase and service sharing, which may include services that may provide location and context sensitive information. The paper discusses number of technical considerations using the proposed framework. These include, the visualization of resources as objects, the novel object oriented model of the mobile grid and the composition of services. The performance of the proposed model is studied using simulation and emulation study.
GDP: A Paradigm for Intertask Communication in Grid Computing through Distributed Pipes
Existing grid models target purely data parallel applications without inter-task communication. This paper proposes a transparent programming model to support communicating parallel tasks in a wide area grid. This paper proposes Grid with Distributed Pipes (GDP), a grid model that enables location independent intertask communication among processes on machines spread over a wide area distributed system. This approach enables anonymous migration of communicating parallel tasks adjusting to grid dynamics. The proposed model supports sequential load to coexist with parallel load. A prototype of the proposed model has been implemented over clusters of nodes spread over the Internet. A steady state equilibrium engineering problem was studied over the model. Performance studies shows linear to super linear speed up for the application.