الفهرس 
Analysis of some Grid ArchitecturesOnly 14 pages are availabe for public view
 |  | from | 32 |  |  |
| | | from | 32 | | |
Abstract
With the rapid advance in science, engineering, and business; people seek more computational power and resources to solve their problems more efficiently in terms of accuracy, time, and money. The field of Grid computing was born to fill the gap between available technology and increasing demand for computational power. The Grid provides a powerful computational environment by coupling distributed resources to enable seamless aggregation and sharing to create more powerful resource. The term distributed here does not refer only to geographical locations but also to administration that may span multiple organizations.
Security issues were addressed from the beginning of the Grid computing because of their importance to the success of such field. Intrusion detection is an important component of any modern security system because it is considered as a second line of defense against bugs and security holes as well as providing protection against insiders.
This thesis studies the problem of intrusion detection in Grid environments since it is considered as an important security issue. It introduces flexible cooperative distributed intrusion detection architecture for computational Grids. This work is based on the study of latest Grid projects and intrusion detection systems to deliver an architecture that suits and benefits form the underlying computational Grid environment.
A prototype implementation of the proposed architecture for the purposes of validation and verification is also introduces. The presented prototype uses homogeneous distributed intrusion detection servers that use the Learning Vector Quantization (LVQ) neural network for classification to detect intrusion cases if occurred ifferent parameters that may affect the proposed intrusion detection The introduced prototype was tested against various Grid environments with different organizations and architectures through a Grid environment simulator that was developed to suit the study of security and intrusion detection. The test results showed the applicability of the proposed system in Grid environments and also showed distinct advances versus centralized systems. The thesis also presents the dsystem showing and explaining their effects on the overall system performance.