Optimization of Replica Consistency and Conflict Resolution in Data Grid Environment

Data replication is widely used mechanism aiming at ensuring efficient access, improving performance and providing quality data in data grids. The properties of file and their replicas can vary with time resulting to inconsistency. In this paper, an agent-based replica consistency approach, namely, Replica Consistency and Conflict Resolution (RCCR) has been proposed to improve the efficiency of the system. RCCR is hybrid approach which is used to address the problem of inconsistency in multi master environment. Additionally, the conflict resolution approach is introduced based on vector clock and version of the replica. Finally, simulated results are presented and compared with existing consistency approaches such as Pessimistic, Optimistic and One-way Replica Consistency (ORCS). The simulated results depicts that proposed RCCR approach performs better when writeable replicas are more likely to occur.

VECTOR CLOCK TRACING AND MODEL BASED PARTITIONING FOR DISTRIBUTED EMBEDDED SYSTEMS

Tracking, partitioning and tracing in modern dynamic high performance computing systems are three of the most innovative and important development aspects for performance optimization purposes and state-of-the-art advanced quality. This paper discusses these three aspects with respect to distributed systems and proposes new mechanisms for an advanced utilization of software in this domain. We present a specific tracking mechanism via vector clocks for model and code partitioning purposes and the determination of causality relations. Further, a tracing approach for an effective analysis and thereby utilization of code and the corresponding architecture is introduced. The combination of both approaches leads to a high degree of parallelism and a fine-grained structure of execution units, that further traced, supports a precise analysis of synchronous and asynchronous system’s behavior as well as an optimal load balancing. The mechanisms are introduced with respect to a model based control engineering tool and event diagrams.