PlanetLab Now and into the Future

1
PlanetLabNow and into the Future

• Marc Fiuczynski, Ph.D.Princeton
  UniversityAssociate Research ScholarPlanetLab
  ConsortiumRD Staff MemberPlanetWorks
  CorporationFounder

2
What is PlanetLab?

• Consortium joint Academic, Government, Industry
  venture
• Formally formed January 2004
• Hosted by Princeton University, UC Berkeley, and
  U. of Washington
• United States Government funded (NSF and DARPA)
• Hewlett Packard and Intel as founding Industrial
  members
• ATT, France Telecom, Polish Telecom, Google,
  NEC,
• Facility Planetary-scale overlay and
  underlay network
• 700 Linux-based servers at 300 sites in 30
  countries
• Researchers can get a virtual machine on each
  server (SLICE)
• In a SLICE across PlanetLab researchers can
  deploy evaluate
• distributed systems services and applications
• The next Internet will be created as an
  overlay in the current one
• network architectures and protocols
• The new Internet will be created in parallel
  next to the current one

3
PlanetLab Facility Today

• 784 servers at over 382 sites
• Co-located throughout the (developed) world _at_
  Uni. Companies
• Co-located at network crossroads (Internet2, RNP,
  CERNET, )

4
The Importance of Systems Building

• Systems-oriented CS research needs to build and
  try out its ideas to be effective
• Paper designs are just idle speculation
• Simulation is only occasionally a substitute
• We need
• Real implementation
• Real experience
• Real network conditions
• Real users
• To live in the future

5
Limitations of Traditional Approaches

• Simulation based on limited models
• Topologies, administrative policies, workloads,
  failures
• Emulation (and in lab tests) are similarly
  limited
• Only as good as the models
• Conventional testbeds are (too narrowly) targeted
• Not cost-effective to test every good idea
• Often of limited reach no real users
• Often with limited programmability

6
Gap between Research Real World
Maturity
DeployedInternet Solution
This gap is a major barrier to introducing novel
technologies Into the Real World Internet!!!
Conventional Testbeds
Emulation-based Prototype
Simulation Analysis
Foundational Research
Time
7
PlanetLab Decreases this Gap
Maturity
DeployedInternet Solution
PLANETLAB
Conventional Testbeds
Emulation-based Prototype
Simulation Analysis
Foundational Research
Time
8
NSF GENI magnifies PlanetLab ideas
Maturity
DeployedInternet Solution
NSF 350M-700M
GENI
PLANETLAB
Conventional Testbeds
Emulation-based Prototype
Simulation Analysis
Foundational Research
Time
9
Long-Running Services

• Content Distribution
• CoDeeN Princeton
• Coral NYU, Stanford
• Cobweb Cornell
• Storage Large File Transfer
• LOCI Tennessee
• CoBlitz Princeton
• Information Plane
• PIER Berkeley, Intel
• PlanetSeer Princeton
• iPlane Washington
• DHT
• Bamboo (OpenDHT) Berkeley, Intel
• Chord (DHash) MIT

10
Services (cont)

• Routing / Mobile Access
• i3 Berkeley
• DHARMA UIUC
• VINI Princeton
• DNS
• CoDNS Princeton
• CoDoNs Cornell
• Multicast
• End System Multicast CMU
• Tmesh Michigan
• Anycast / Location Service
• Meridian Cornell
• Oasis NYU

11
Services (cont)

• Internet Measurement
• ScriptRoute Washington, Maryland
• Pub-Sub
• Corona Cornell
• Email
• ePost Rice
• Management Services
• Stork (environment service) Arizona
• Emulab (provisioning service) Utah
• Sirius (brokerage service) Georgia
• CoMon (monitoring service) Princeton
• PlanetFlow (auditing service) Princeton
• SWORD (discovery service) Berkeley, UCSD

12
Usage Stats

• Slices 600
• Users 2500
• Bytes-per-day 4 TB
• IP-flows-per-day 190M
• Unique IP-addrs-per-day 1M

13
Three Key Ideas in PlanetLab

• Virtualization (Slices)
• Supports concurrent research on shared facility
• Amortizes the cost of building the facility
• Enables long-running experiments and services
• Programmable
• Enable prototyping and evaluation of new services
• Enable a revisiting of todays protocol layers
• Opt-in on a per-user / per-application basis
• Attract real users
• Demand drives deployment / adoption
• Connect to the Internet
• To reach users, and to connect to existing
  services

14
PlanetLab Node Software Architecture
Slice Manager (SM)
Slice
Slice
Slice
Slice
15
PlanetLab Management Architecture
Global Slice Management
PLC
Resource Controller
Auditing Archive
node control
sensor data
16
Slices
17
Slices
18
Slices
19
User Opt-in
Client
Proxy
20
PlanetLab Growth 07
Russia
More Brazilian Universities?! Latin America
(RedClara)
Japan (JGN2)
21
PlanetLab (up until) Today
Utility Computing Utopia
1000s
Compute Farms
100
10
Nodes per Site
PlanetLab
1
p2p
routers
106s
1
10
100
1000s
Sites
22
PlanetLab Future
Utility Computing Utopia
1000s
Compute Farms
100
10
Nodes per Site
PlanetLab
1
p2p
routers
106s
1
10
100
1000s
Sites
23
Other PlanetLab based Testbeds

• PlanetLab EU PlanetLab JP
• Federated with PlanetLab
• EverLab
• private PlanetLab for HPC and Networking
  reseachers (EU Israel)
• OneLab Intels Wireless PlanetLab
• PlanetLab for wireless research (France)
• VINI
• PlanetLab for layer 3 research

24
VINI PlanetLab for layer 3 research
s
25
Summary

• PlanetLab provides a blueprint for introducing
  novel technologies into the Internet
• More information / contacts
• PlanetLab http//www.planet-lab.org
• mef_at_planet-lab.org
• mef_at_cs.princeton.edu
• GENI http//www.geni.net