To provide intelligent solutions for computational/data intensive applications in the distributed environment.
To establish a world-class research centre to produce human resources in the area of Grid Computing.
This project addresses various limitations of current globus toolkit in security and highlights the importance of additional security and trust features in globus toolkit. Currently, most of the grid middle-wares including Globus Toolkit use the existing X509 authentication mechanism. However, this mechanism does not guarantee for complete secure network and cannot carry dynamic data. Also, it is rigid, non-scalable, highly limited to organization grids. In this project, we proposed and implemented Kerberos based authentication architecture for Enterprise Grids to authenticate users from an organization or from outside organization. This proposed architecture addresses the issue of Security requirements for Enterprise grid services. Further, the architecture does not require Kerberos client software in the client side and is implemented as grid service that supports Service Oriented Architecture. This portable nature of the authentication mechanism proposed in this project facilitates the use of grid environment for commercial grid services and also for utility computing. Also, In Enterprise Computational Grid environment, the computation in the application needs to process large number of data and requires more number of processor cycles such as cryptographic system, disaster management system etc. Hence, it is mandatory to determine the correctness of the result obtained and hence the service provider.
In conventional approach, Merkle Tree is used to calculate the honestness of the service providers and it addresses the issue of sample selection. However, this approach does not provide any efficient way for selecting the samples and hence it is not applicable to compute the trustworthiness of the service provider. The main motivation behind this research work is that, the calculation of trustworthiness needs some enhancement to the conventional approach. As a part of this project work, we implemented a methodology that assigns weightages to the samples depending on their complexity so that it enables one to select various samples while determining honestness of the computation. We apply this approach to evaluate the trustworthiness of a service provider. Also, we prove that the proposed approach is more efficient than the conventional approach using suitable illustrations. The project also addresses the complexity involved in choosing appropriate parameters for calculation of trust component. Trust plays an important role in all commercial grid interactions. One of the major challenges for commercial grid systems is how to establish a trust relationship between different parties. Trust component is the central component of this project.
In this project, we propose Multi-dimensional Trust model which is based on parameters such as service, Transaction, Technology, Institutional, Social and Consumer Behavioral dimensions. Each dimension consists of various sub-dimensions to reflect the overall Trustworthiness (T) of a given system. We mathematically derived an equation on each dimensions. This framework suit in evaluating the trustworthiness of various entities viz Service provides, Customers, Trust Manager in commercial applications. We are in the verge of designing a new evolutionary job scheduler that calculates trustworthiness of the party involved in commercial grid environment for service access and schedules the job based on the trust value.
Globus Toolkit 4.x, Apache Jakarta Tomcat, Gridsphere Framework, JDK1.5,J2EE, JSP, MPICH-G2, Beowolf Cluster configuration, Various Components of Globus Toolkit including GSI, SimpleCA, GridFTP, MDS, GRAM,RFTP, RLS, Condor-G, PBS and Moab scheduler, XIO, Triana Workflow, Permis Authorization toolkit, Kerberos Authentication Mechanism, Oracle 10G