1 / 4

Revisiting DB Server Basics: Modernizing APIs & Embracing Middleware for Enhanced Performance

Delve into the evolution of DB server products, challenge 1980s assumptions, and explore solutions for modern requirements like lightweight APIs, paginated results, and better integration with middleware for optimal performance.

rollandd
Download Presentation

Revisiting DB Server Basics: Modernizing APIs & Embracing Middleware for Enhanced Performance

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. Going, Going, … Gong!(a.k.a. What I Wish DatabaseResearchers Would Solve For Me) Mike Carey BEA Systems, Inc. May 2003

  2. Time to Revisit DB Server Basics! • Basic design of today’s major DB server products dates back to the Client-Server Era (~early 1980’s) • Revisit server architectures and their APIs! • 1980’s assumptions no longer hold (if they ever did) • Clients don’t talk to DB servers – applications talk to DB servers (ex: SAP, J2EE apps, web services) • Application servers and applications end up redoing and/or working around DBMS features • Security ends up in the application • App servers or apps do connection pooling • Apps use programming tricks / patterns to deal with disconnection, result pagination, etc.

  3. Idea #1: Modernize DBMS APIs • Revisit client/server state management • Everything must be lighter and cheaper! • Connections (no pooling) • User identity management • Fix the client/server result handling model • Lightweight paginated results • Ex: 1003 matching products – first 20, next 20, … • Let client consume query results incrementally • Don’t hold critical DB resources (locks, buffers, …) • Optimistic updates (single- and multi-row) • Read / think / update support for interactive web apps (e.g., disconnected rowsets) is ad hoc • Now time to ignore [Agrawal, Carey, Livny 1987]? ()

  4. Idea #2: Be Middleware-Friendly • DB servers ought to cooperate with middleware that does caching and needs to be change-aware • EJB servers, data integration servers, fancy apps • Note: Remember that the world is heterogeneous! • I’m not sure what this means, but it might mean • Support gazillions of triggers, and/or • Add external change eventstream APIs, and/or • Add native (standard) support for version stamps?

More Related