PhD Thesis: Security and Management in Networked Storage Services
Storage has traditionally been offered as an intrinsic component of a computing environment - being directly attached to servers and PCs. Even with the advent of networked storage technologies like Storage Area Networks (SAN), the storage system has been contained within the organizational boundaries of an enterprise. An alternative paradigm to this intrinsic storage is to provide storage as a service over the internet. In this utility-based model, storage can be procured and used on an on-demand basis. This paradigm is motivated by many compelling reasons and is being increasingly adopted in both enterprise storage systems (as evident by the growth of Storage Service Providers - SSPs) as well as consumer storage systems (online storage, Peer-to-Peer storage).
The figure above shows an example enterprise storage-as-a-service environment. By storing the data with a dedicated SSP, the enterprise avoids the expensive and complex task of storage management, can better adapt its storage backend to its data growth, ensure a superior disaster recovery (DR) plan and better regulatory compliance strategy, e.g. storing data on a Write-Once-Read-Many (WORM) media.
This new outsourced storage environment presents many challenging problems, the foremost being that of usable security and privacy techniques. As the storage device and its data reside externally at service providers that might not be fully trusted, it is of paramount importance that this data is stored and accessed in a reliable and secure manner. The second challenge is the management of service provider's storage infrastructure, which has to deal with a highly dynamic environement with large data capacity requirements and distributed data generation and access.
My dissertation research addresses these two challenges in depth. First, we analyze the new security and privacy requirements of outsourced enterprise storage systems for different service models. We develop techniques for data privacy enhancement and propose effective mechanisms for secure data storage and flexible access control [1, 2, 3]. We also propose a secure multiuser file system search [4] that allows service-provider to host search indices over archived data. All developed models and techniques are designed with an emphasis on usability and user convenience and do not require the existence of a trusted service provider. Second, we tackle the storage management challenge for storage service providers by designing efficient change management techniques for storage area networks [5], dynamics handling in virtualized data center environments [6] and a storage outsourcing planner [7].
- SHAROES: A Data Sharing Platform for Outsourced Enterprise Storage Environments, Aameek Singh, Ling Liu, IEEE International Conference on Data Engineering (ICDE) 2008. [pdf][ppt]
- Privacy Analysis and Enhancements for Data Sharing in *nix Systems, Aameek Singh, Ling Liu, Mustaque Ahamad, to appear in International Journal of Information and Computer Security (IJICS), 2008.
.
- Privacy Analysis for Data Sharing in *nix Systems, Aameek Singh, Ling Liu, Mustaque Ahamad, USENIX Annual Technical Conference 2006. [pdf][ppt]
- Efficient and Secure Search of Enterprise File Systems, Aameek Singh, Mudhakar Srivatsa, Ling Liu, IEEE International Conference on Web Services (ICWS), 2007. [pdf][ppt]
- Zodiac: Efficient Impact Analysis for Storage Area Networks, Aameek Singh, M. Korupolu, K. Voruganti, USENIX File and Storage Technologies (FAST) 2005. [pdf][ppt]
- SPARK: Integrated Resource Allocation in Virtualization-Enabled SAN Data Centers, Aameek Singh, Madhukar Korupolu, Bhuvan Bamba, IBM Research Report RJ10407, submitted. [pdf][ppt]
- BRAHMA: Planning Tool for Providing Storage Management as a Service, Sandeep Uttamchandani, Kaladhar Voruganti, Ramani Routray, Li Yin, Aameek Singh, Benji Yolken, IEEE Int'l Conference on Services Computing (SCC) 2007. [pdf][ppt]