1 / 9

Storage management and caching in PAST, a large-scale, persistent peer-to-peer storage utility

Storage management and caching in PAST, a large-scale, persistent peer-to-peer storage utility. Antony Rowstron, Peter Druschel. Introduction. PAST is an Internet-based, peer-to-peer global storage utility. It aims to provide strong persistence, high availability, scalability and security.

Download Presentation

Storage management and caching in PAST, a large-scale, persistent peer-to-peer storage utility

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Storage management and caching in PAST, a large-scale, persistent peer-to-peer storage utility Antony Rowstron, Peter Druschel

  2. Introduction • PAST is an Internet-based, peer-to-peer global storage utility. • It aims to provide strong persistence, high availability, scalability and security. • PAST employs Pastry as the routing scheme. • Pastry is an efficient routing scheme used to reliably route client requests to the appropriate nodes.

  3. PAST Operations • fileId=Insert(name, owner-credentials, k, file) fileId is 160-bit, while nodeId is 128-bit; fileId nodeId • file=Lookup(fileId) Reliably retrieve a copy of the file • Reclaim(fileId, owner-credentials) Notice: Reclaim doesn’t guarantee that the file is no longer available after it was reclaimed. 128 Bits 32 Bits M L 128 Bits

  4. Storage Management • Design Goals: • Allow high global storage utilization • Gracefully degrade as the system approaches the maximal utilization • Load balancing: • Replica diversion • File diversion

  5. Replica Diversion • Replica diversion is to accommodate differences in the storage capacity and utilization of nodes within a leaf set. File is stored in B due to the less capacity in A B C C is the k+1th closest node, which stores the pointer in case of the failure of node A An entry points to B A A is the right node to store the file, but has insufficient space

  6. File Diversion • The k closest nodes to the fileId could not accommodate the file nor divert the replicas; then, a file diversion is invoked. • The client generates a new fileId, and retries to insert the file. • Repeat the second step up to 3 times. • If it still fails, an error is reported to the application.

  7. Caching • PAST nodes use the “unused” portion of their advertised disk space to cache files. • Caching in PAST is based on GreedyDual-Size (GD-S) policy, which was originally developed for caching Web proxies.

  8. Experimental Results *This graph is from the paper.

  9. Question?

More Related