1. Grid Computing, Database Architecture for GEMLCA Resource Availability

1.1 Introduction

Computational Grids connect resources and users in a complex way in order to deliver nontrivial qualities of services. The definition of grid computing stems from the idea of providing a generic computational infrastructure, completely transparent to its users, capable of solving a wide set of problems which may be related, but not limited to information collection, elaboration and storage.
Each user accesses Grid services with the expectation of the resources required to be available and accessible in order to execute workflows that overcome the difficulties of application development and execution more easily. However users cannot be guaranteed that the selected resources where the execution of workflow components is deployed are available and functional. The problem domain this project covers is a new workflow-oriented portal concept that is aimed at the GEMLCA (Grid Execution Management for legacy code architecture) to analyse and accurately provide information on resource availability and a probability analysis of resources being available or  live at the time of workflow execution. This enables the various Grid clients to be mapped to the functional Grid sites in order to submit their workflow execution
This functionality will be introduced through the provision of a historical database model that can be used to identify, collect and collate data for workflow-oriented Grid portals based on two
general features: Ability to access multiple Grids, and support for collaborative problem solving by provision of a probability analysis of a resource that may not be available at the time of the work flow execution, but will be imminently available within a specified timescale.
The provision of this high-level information to the grid client is extremely crucial and vitally important in order to determine where workflow execution can be remapped imminently or provide a dynamically verified resource. This data will be stored in a historical database and can be accessed by the GMT classifier and production Grid brokers and by utilising extensible algorithms, behavioural patterns for analysing chronological occurrences can be constructed for specific resource availability and calculate probabilities of a resource or service being functional within a specific timescale.
The Grid is a highly distributed set of technologically different and
generally transient computing resources, we can easily identify a range of problems
related to the dynamic discovery of resources together with the availability of information
regarding the state and current capabilities of these resources. To allow for real user
transparency, resource discovery and brokering, the grid must be capable of adapting to
particular temporary conditions such as network partitioning, computational nodes
overload and data storage unavailability and the availability of a historical database is a versatile method of predicting resource allocation to all the different sites.
Harvard style referencing up to 50 references required.
plese see this portlet