Uniform TitleA decentralized content-based communication framework for supporting decoupled grid interactions
NameQuiroz Hernandez, Andres (author), Parashar, Manish (chair), Marsic, Ivan (internal member), Silver, Deborah (internal member), Rutgers University, Graduate School-New Brunswick,
SubjectElectrical and Computer Engineering,
Computational grids (Computer systems)
DescriptionThis work presents a decentralized and content-based communication framework based on Web Services Notification (WSN) to enable monitoring and loosely-coupled interactions in distributed and particularly Grid systems. Web services have emerged as one of the key enabling technologies for Grid systems, providing platform-independent interactions between distributed applications and resources. The WSN specification is a set of web service standards that define protocols for realizing the publish/subscribe communication pattern. Existing implementations of WSN fail to address key issues particular to large-scale Grid systems, such as the dynamic participation of distributed computing nodes, the heterogeneous services provided by nodes across physical and organizational domains, and the need for efficient messaging mechanisms.
The specific contributions of this work are as follows: 1) Design and development of a notification service implementing the WSN specifications based on a decentralized content-based addressing and messaging infrastructure; 2) Design and evaluation of self-optimization mechanisms for message exchanges within the notification system; 3) Design and analysis of a novel two-level overlay structure that extends the messaging infrastructure for efficiently interconnecting separate individual physical or organizational groups of computation nodes without resorting to designated nodes that can become bottlenecks or single points of failure; 4) Development of a mechanism for self-monitoring of peer-to-peer systems that is able to detect anomalies and/or trends in system behavior as a direct application of the notification service.
The experimental evaluations presented demonstrate the performance and scalability of the infrastructure as well as the effectiveness of the self-monitoring mechanism.
Note[bibliography] Includes bibliographical references (p. 68-71).
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.