View Full Document

Two-Level, Self-Verifying Data for Peer-to-Peer Storage



View the full content.
View Full Document
View Full Document

3 views

Unformatted text preview:

Two Level Self Verifying Data for Peer to Peer Storage Patrick Eaton Hakim Weatherspoon and John Kubiatowicz University of California Berkeley Report No UCB CSD 05 1401 June 2005 Computer Science Division EECS University of California Berkeley California 94720 Two Level Self Verifying Data for Peer to Peer Storage Patrick Eaton Hakim Weatherspoon and John Kubiatowicz University of California Berkeley June 2005 Abstract tructure then manages indexes tracks and repairs each small block individually First generation peer to peer storage systems unnecessarily couple the unit of client data access to the unit of infrastructure data management Designs that require all peers to operate on data at a fixed granularity lead to inefficiencies such as high query load and high per block storage overheads To provide variable granularity access and support more efficient peer to peer storage systems we introduce two level naming of self verifying data We describe how to implement two level naming and advocate an extension to the traditional API used by peer topeer storage systems to support two level naming Such a design effectively couples the infrastructure s unit of data management to the client s unit of data creation and access This coupling engenders a challenge for creating efficient systems For clients at the edges of the network it is natural to work with data divided into small blocks However a storage infrastructure that manages small blocks sees higher indexing cost because the infrastructure must track all replicas of each smaller block individually and a larger query load because clients must use the infrastructure to locate each small block On the other hand a storage infrastructure can reduce the overhead of tracking data and resolving queries by amortizing the costs over larger blocks of data However requiring clients to work with large blocks can waste precious bandwidth at the edges of the network and require significant data buffering that limits data



Access the best Study Guides, Lecture Notes and Practice Exams

Loading Unlocking...
Login

Join to view Two-Level, Self-Verifying Data for Peer-to-Peer Storage and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Two-Level, Self-Verifying Data for Peer-to-Peer Storage and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?