1 / 10

Khazana: Flexible Wide Area Consistency Management

Khazana: Flexible Wide Area Consistency Management . Sai Susarla John Carter School of Computing University of Utah. Wide Area Consistency Management. Wide area services are not alike Different consistency and availability requirements Ideal: Reuse mechanisms from existing systems

thao
Download Presentation

Khazana: Flexible Wide Area Consistency Management

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. Khazana: Flexible Wide Area Consistency Management Sai Susarla John Carter School of Computing University of Utah

  2. Wide Area Consistency Management Wide area services are not alike • Different consistency and availability requirements • Ideal: Reuse mechanisms from existing systems • Problem: Each system builds monolithic solution suited only to its own particular needs Khazana: Configurable caching infrastructure • Support mix of caching & availability mechanisms • When is access allowed to a replica (optimistic/pessimistic)? • Where are updates issued (single/multiple masters)? • How are updates propagated (push- or pull-based)? • What causes updates (timer, staleness, eager, lazy)? • Let application tune how its data is managed

  3. Exploit Commonality System files, static web content Pull updates Strict pessimistic Single-master Scoreboard, Active Dir. schema config files, multimedia streaming Push updates Optimistic last-writer-wins, Bayou, Coda, calendar, sensor logging, mobile file access Pull updates Append-only Multi-master Push updates Active Directory data, chat, whiteboard Current Approach: • Implement one combination effectively to suit one app. • Restricts reusability Khazana Approach: • Implement common mechanisms in app-independent manner • Provide hooks to let applications customize their behavior

  4. Khazana Approach • Wide-area peer-to-peer data store • File-like data abstraction, page-grained consistency • Performs caching, concurrency control • Asynchronous update notifications • Application controls per-file behavior • Consistency policy (last-writer-wins, append-only, or strict) • Control over per-replica quality • Optimistic vs. pessimistic access • Push vs. pull updates, single vs. multiple masters • Per-file dynamic cache hierarchy

  5. Replicated Objects in Khazana . . . P S1 S2 S3 P: Primary replica Sn: Khazana server S4 S5 P . . . Object O1: multi-master, pull-updates optimistic consistency (e.g., shared file) Object O2: single-master, push-updates strict consistency (e.g., shared password db) Parent-child relationship between object replicas Client accesses replica

  6. Replicated Objects in Khazana . . . S2 Crashes P S1 S2 S3 P: Primary replica Sn: Khazana server Link down O2 copy inaccessible Broken link S4 S5 P . . . Object O1: multi-master, pull-updates optimistic consistency (e.g., shared file) Object O2: single-master, push-updates strict consistency (e.g., shared password db) Parent-child relationship between object replicas Client accesses replica

  7. Replicated Objects in Khazana . . . P S1 S2 S3 P: Primary replica Sn: Khazana server Link Up Broken link S4 S5 P . . . Object O1: multi-master, pull-updates optimistic consistency (e.g., shared file) Object O2: single-master, push-updates strict consistency (e.g., shared password db) Parent-child relationship between object replicas Client accesses replica

  8. Replicated Objects in Khazana . . . S2 Recovers P S1 S2 S3 P: Primary replica Sn: Khazana server Broken link S4 S5 P . . . Object O1: multi-master, pull-updates optimistic consistency (e.g., shared file) Object O2: single-master, push-updates strict consistency (e.g., shared password db) Parent-child relationship between object replicas Client accesses replica

  9. A Khazana-based Chat Room 3 callback(handle, newdata) { display(newdata); } main() { handle = kh_open(kid, "a+"); kh_snoop(handle, callback); while (! done) { read(&newdata); display(newdata); kh_write(handle, newdata); } kh_close(handle); } P 4 2 1 Sample Chat client code Chat transcript: multi-master, push updates, optimistic append-only consistency Update propagation path

  10. Status • Prototype WAN implementation working • Supports last-writer-wins policy • Supports choice of • single- vs. multi-master updates, • push vs. pull-based transfer • Target wide-area services • DataStations:distributed file store that supports multiple file sharing patterns (implemented) • Chat room: concurrent appends to a transcript file • Scoreboard: broadcasts game state to registered listeners • Directory: Hash table implemented in a shared file

More Related