Vidya Ranganathan

Object Isolation for Cloud with DOMAIN RBAC

Cloud computing has taken technology into a mix of networking, virtualization and clustering environment which has opened up a new era with lots of opportunities thereby making business highly scalable. However there are several challenges that need to be addressed, in particular, security, which Forrester [1] has listed as being one of the most crucial concerns. One of the most effective and time-tested ways to ensure security is via Role Based Access Control (RBAC) [2]; with emphasis in cloud computing environments on data protection, authentication and authorization. RBAC provides a policy framework to enable delegation of responsibilities of the super user permissions to other users. This framework helps define non-root users with proper authorizations to perform specific system administration tasks. However it does not provide a mechanism to define the set of objects on which these roles could be exercised. By default, all Role based tasks can be performed on all objects of that type. Therefore to address this issue DOMAIN RBAC has been implemented, with object isolation feature included, as an extension of RBAC. Object Isolation marks a boundary across system resources and users by defining which users can access specified resources on the system while the RBAC roles would determine what operations can be performed on the accessible resources. In this paper we present and describe a method by which object isolation has been implemented via DOMAIN RBAC along with a use case. We also illustrate our approach by showing how it is implemented on IBMs AIX version 7 Operating System which can be leveraged in cloud environment.

Performance Boost Framework in UNIX Kernel for Cloud Applications

Volcanic growth of expanding business and increasing market dimensions lead to need for high performing applications particularly in this internet era. Forrester predicts performance and availability should not to be taken lightly. To fight this onslaught, a business organisation no longer can just accelerate current practices and rely on human intelligence. The companys software also needs to be autonomic. One key aspect for a software to be considered as autonomic is its ability to analyse the system conditions and invoke appropriate actions to optimise the performance and thereby maintaining user satisfaction. Gone are those days where a system would have about five or ten features embedded into the kernel which would have about few hundred services to serve the application requirements. With growing data center it will be impossible to maintain the performance of an application or a system one by one. This clearly dictates the need for automation. In this paper I will discuss about a performance boost framework in UNIX kernel which shall enable the applications take appropriate actions immediately. I shall also provide a demonstration of how this is achieved in a UNIX platform and deployed into distributed cloud environment.