A Distributed Key Based Security Framework for Private Clouds

Abstract

—Cloud computing in its various forms continues to grow in popularity as organizations of all sizes seek to capitalize on the cloud's scalability, externalization of infrastructure and administration and generally reduced application deployment costs. But while the attractiveness of these public cloud services is obvious, the ability to capitalize on these benefits is significantly limited for those organization requiring high levels of data security. It is often difficult if not impossible from a legal or regulatory perspective for government agencies or health services organizations for instance to use these cloud services given their many documented data security issues. As a middle ground between the benefits and security concerns of public clouds, hybrid clouds have emerged as an attractive alternative; limiting access, conceptually, to users within an organization or within a specific subset of users within an organization. Private clouds being significant options in hybrid clouds, however, are still susceptible to security vulnerabilities, a fact which points to the necessity of security frameworks capable of addressing these issues. In this paper we introduce the Treasure Island Security Framework (TISF), a conceptual security framework designed to specifically address the security needs of private clouds. We have based our framework on a Distributed Key and Sequentially Addressing Distributed file system (DKASA); itself borrowing heavily from the Google File System and Hadoop. Our approach utilizes a distributed key methodology combined with sequential chunk addressing and dynamic reconstruction of metadata to produce a more secure private cloud. The goal of this work is not to evaluate framework from an operational perspective but to instead provide the conceptual underpinning for the TISF. Experimental findings from our evaluation of the framework within a pilot project will be provided in a subsequent work.