DATA ARCHITECTURE REPLICATION SERVICE - OGSA

This article is for analyzing the elements that make up a data replication service based on the data architecture model proposed by OGSA. The collection of the information was carried out exhaustively, reviewing the material contained in the books specialized in Distributed Systems, magazines and the Internet. The replication service allows the movement of data in the grid, transparently and dynamically through the creation of local copies, which can be updated or deleted. The analysis covers two types of models: the generic model composed of a primary data source, the replicas and the replication service itself, and a second model implemented in a peer-to-peer environment. Operations such as modifying the data contained in a replica, verifying the consistency of replica data, and managing replica data are analyzed. The user needs to feel supported, to perform their operations, but the instant a replica is updated, the others are outdated. It is important to determine the impact of this outdated on the grid and reduce it through the use of different consistency models, which can include the value by its which replicas are different, going through the calculation of time, up to a maximum number of attemptative writes to be performed. Upgrading replicas is a challenge for the grid system.