USITS '03 Keynote Address
PlanetLab: an open community test-bed for Planetary-Scale Services
a work in progress, 3/26/03
David E. Culler, University of California, Berkeley, and Intel Research
Click here to start
Table of Contents
- PlanetLab: an open community test-bed for Planetary-Scale Services
- PlanetLab today
- Where did it come from?
- key missing element hands on experience
- Guidelines (1)
- Guidelines (2)
- Guidelines (3)
- Guidelines (4)
- Guidelines (5)
- Confluence of Technologies
- March 02 “Underground Meeting”
- Outcome
- Tension of Dual Roles
- Overlapping Phases
- Architecture principles
- Slice-ability
- Unbundled Management
- Distributed Resource Control
- Application-Centric Interfaces
- Service-Centric Virtualization
- Changing VM landscape
- How much to virtualize?
- Slivers of a Slice: long-term plan
- Kickoff to catalyze community
- the meta-testbed effect
- Current Approach (on to phase I)
- vServer experience (Brent Chun)
- SILK (Princeton)
- Dynamic Slice Creation
- BootCD enabling growth
- A typical day (1/28)
- Run up to SIGCOMM
- A Slice for a Month (Duke)
- So what are people doing?
- Really...
- with the internet in the middle
- Analysis of Tapestry (Ben Zhao)
- Towards an instrumentation service
- Ossified or fragile?
- the Gaetano advice
- ScriptRoute (Spring, Wetherall, Anderson)
- Example: reverse trace
- Ossified or brittle?
- NetBait Serendipity
- One example
- No, not Iraq
- Netbait view of March
- DHT Bakeoff
- Rationalizing Structured P2P Overlays
- Empirical Comparison (Rhea, Roscoe,Kubi)
- Bake-off Controversy
- Distributed Storage
- Dipping in to OceanStore Prototype
- Tenn IBP traffic
- SOSP Crunch
- Deep resource management
- Example Services
- Evolving the Architecture (Peterson)
- What Planet-Lab is about
- Join the fun ... www.planet-lab.org
- Thanks
|