1 / 5

Tapestry: Wide-area Location and Routing

Tapestry: Wide-area Location and Routing. Ben Y. Zhao John Kubiatowicz Anthony D. Joseph U. C. Berkeley. Why Tapestry?. Distributed systems scaling to WAN Larger scale  frequent component faults More data + centralization  performance bottleneck

donnaaroberts
Télécharger la présentation

Tapestry: Wide-area Location and Routing

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. Tapestry: Wide-area Location and Routing Ben Y. Zhao John Kubiatowicz Anthony D. Joseph U. C. Berkeley

  2. Why Tapestry? • Distributed systems scaling to WAN • Larger scale  frequent component faults • More data + centralization  performance bottleneck • Dynamic environment  manageability complexity • More principals  attacks on system (e.g. DoS) more likely • Tapestry: • Decentralized approach to location and routing focusing on fault-resilience and adaptability • Builds on previous work: Plaxton trees OSDI 2000 WiP

  3. Wide-area naming Nodes/Objs named by hashed bit-sequence IDs Incremental routing Route to root via local neighbor maps Incremental progress towards destination Properties Exploits search locality Route around failures Decentralized scaling LogbN hops to destination 3B8C 785A 8963 0A25 613A 903A 12D2 203A Plaxton Trees Route(3B8C->203A) RoutingMap(abcd) OSDI 2000 WiP

  4. Tapestry Improvements • Root nodes => single point of failure • Soln: Root redundancy via hash salts • Topology changes => high cost • Soln: Local heartbeats, alternate pointers, second chance invalidation • Dynamic system => error persistence • Soln: Proactive node-integration, fault-detection, Self-optimization via query state • Vulnerable to DoS attack • Soln: Approx. nodes for load diversion, online data verification, compromised node isolation OSDI 2000 WiP

  5. Project Status • Providing location/routing support for the Oceanstore global storage project • http://oceanstore.cs.berkeley.edu • Java-based prototype • C-based simulation / measurements • For more details, see Poster Session • Contact • ravenben@cs.berkeley.edu OSDI 2000 WiP

More Related