1 / 22

Kinesthetics eXtreme - KX

Kinesthetics eXtreme - KX. Scenario: Distributed system where everything is running smoothly. Oops. raw events. Let’s probe into this. How is this done?. Active Interfaces – before/after callback model for probe insertion

frieda
Télécharger la présentation

Kinesthetics eXtreme - KX

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. Kinesthetics eXtreme - KX

  2. Scenario: Distributed system where everything is running smoothly

  3. Oops

  4. raw events Let’s probe into this

  5. How is this done? • Active Interfaces – before/after callback model for probe insertion • AIDE (Active Interface Development Environment)– tool for automatically instrumenting Java source code • In collaboration with Prof. George Heineman at WPI

  6. Smart Events Package things up raw events

  7. How is this done? • FleXML (Flexible XML) – smart event representation extends XML, splits lightweight meta-data from heavyweight content • Event Packager (aka Flight Recorder) – converts raw events from probes to FleXML smart events using probe-specific tagger plugins, logs for “latecomers” • Siena (from University of Colorado at Boulder) – Internet-scale event bus filters and routes “flattened” FleXML smart events as sequences of attribute/value pairs

  8. Check things out raw events Smart Events

  9. How is this done? • Event Distiller – subscribes to Packagers and other Distillers producing smart events that may partially match complex event patterns over partially ordered multi-source event streams • FleXML Meta-Parser – validates smart events against schema fragments, invokes tag processors on behalf of Distillers

  10. Hmm… this doesn’t look right raw events Smart Events

  11. How is this done? • FleXML Oracle – users install new schema fragments and tag processors here, lazily deploys to Meta-Parsers as they receive previously “unknown” or “invalid” tags • Workgroup Cache – eagerly deploys FleXML schemas/processors to related Meta-Parsers, “group”-based prefetch/cache/push rules also support information sharing within and among human teams • Worklets - mobile agents that carry and install probe callbacks, tagger plugins, event patterns, schema fragments, tag processors, etc.

  12. Smart Events Get the word out raw events Smart Events

  13. How is this done? • Event Notifier – subscribes to Distillers producing smart events that may trigger “gauge” updates • FleXML Meta-Parser – validates smart events against schema fragments, invokes tag processors on behalf of Notifiers

  14. Gauging what’s up raw events Smart Events Smart Events

  15. How is this done? • Gauges - render smart events into human-visible graphics, reformat smart events for external analysis tools • TRIKX (TRansitional Interface for KX) – portal organizes user or role-specific portlets for system management and gauge presentation to human and/or automated decision agents

  16. Send out the gaugents raw events Smart Events Smart Events

  17. Reconfigure raw events Smart Events Smart Events

  18. How is this done? • Process-aware systems employ decentralized workflow to semi-automatically instantiate and coordinate worklets • Other human and/or automated approaches for developing and managing worklets in general and gaugents in particular • Gaugents – specialized worklets oriented to system reconfiguration and repair, require Worklet Virtual Machine (WVM) adaptor for each affected target system component

  19. Yes! !

  20. Challenges • CHIME 3 • Apply KX to monitoringbackend data changes, morphing frontend client world views for live users, and interpolating contextualized frontend client world views for offline and latecomer users • Adaptive Interactive Internet Team Video (AI2TV) joint with Profs. John Kender and Jason Nieh • Apply KX to contextualized prefetching and variable-bandwidth synchronizing of streaming video depicted within CHIME virtual worlds targeted to team-oriented distance learning

  21. Extend FleXML to heavyweight information objects (e.g., streaming video, large images) Complex event pattern notations/recognizers addressing distributed computing protocols and temporal/causal relationships Prefetching, pushing and contextualization models and algorithms Decentralized workflow management system (e.g., based on peer file sharing models and infrastructure) Portal Framework based on GC-4 (Groupspace Controller version 4) vetoable events and service role brokering WHEATIES (Wimpy HandhEld-Accessible Tracking, Identification and Error correction System) – KX for PDAs Projects

More Related