A secure grid computing model and its application in DNA database search

Grid computing is an ad hoc form of network computing used particularly in science and engineering computing applications. It is distinguished from traditional distributed computing by its large scale of resource usage across multiple administrative domains which is attractive to computation-intensive and data-intensive projects in many fields.; Secure Grid computing involves addressing security issues in heterogeneous wide-area computing environments. Providing flexible security is important in Grid architectures to meet different demands in applications working in distinct underlying systems. Much of the current focus in Grid security research has been in the area of deterministic symmetric cryptography. However, probabilistic cryptosystems have a number of advantages in that they provide semantic security to sensitive data transfer across the Grid platforms. Furthermore, using dense probabilistic encryption (DPE) reduces the encryption overhead.; The objectives of this thesis are two-fold. First, a secure Grid computing model is designed providing message confidentiality based on block-based DPE, mutual authentication, data integrity and digital signature. Second, the application of the model in DNA sequence database search is demonstrated. The security features are tested in a virtual Grid environment along with an implementation of the application of similarity comparisons in DNA databases. Modification is made to the sensitive similarity comparison and local alignment algorithm, namely, the Smith-Waterman (SW) algorithm, to reduce its complexity.; Experimental results show that the DNA sequence comparisons are executed efficiently with the modified Smith-Waterman algorithm, and the block-based sequential DPE encryption and parallel decryption provide an efficient way for computation-intensive applications in a secure Grid.