Cell approximation method in quorum systems for minimizing access time

As part of a service oriented e-Science and Grid infrastructure, a quorum-consensus method is a valuable tool for managing replicated data and finding meaningful patterns of immense data sets. In general, a workflow is defined as set of atomic tasks that may have dependencies between each other. Tasks also may define a so-called “quorum” which denotes a number of precedence constraints to be fulfilled before execution may proceed. To solve the problem of designing a quorum that fits a scalable and dynamic environment where data are distributed sparsely or too far apart, we propose a cell approximation method to fit into some form (e.g., voting, grid) of logical structure in quorum-based protocols. In our method, adjacent nodes are configured by an Intra cell which can improve their data locality. Furthermore, Intra-cell relationships can be represented by an Inter cell for minimizing access time.