Information Services and Process Innovation Lab

We have been pursuing techniques, applications and systems for content-addressable storage. We have applied these in a variety of settings. For example, we developed a synchronization system that was used for in the SE3D Film Festival in 2005 and 2006 to enable small animation companies to use HP's utility computing infrastructure and state-of-the-art data centers to render their movies (even though the data centers were an ocean away over public networks and slow and fragile connections). We also actively work with HP's StorageWorks division on a variety of archival, backup, and information-aware storage products.

Our research in this area focuses on better chunking algorithms, index partitioning, robust and efficient synchronization algorithms, and distributed data management.

Selected Publications:

• Jumbo Store: Providing Efficient Incremental Upload and Versioning for a Utility Rendering Service. Kave Eshghi, Mark Lillibridge, Lawrence Wilcock, Guillaume Belrose, and Rycharde Hawkes. 5th USENIX Conference on File and Storage Technologies (FAST '07). HPL-2006-144

• A Framework for Analyzing and Improving Content-Based Chunking Algorithms. Kave Eshghi and Hsiu-Khuern Tang. HPL-2005-30R1

• Intrinsic References in Distributed Systems. Kave Eshghi. HPL-2002-32