The Future of Internet Research - Electrical Engineering & Computer

Download Report

Transcript The Future of Internet Research - Electrical Engineering & Computer

The Future of Internet Research
Scott Shenker
(on behalf of many networking collaborators)
Why Not “The Future of the Internet”?
• We always miss the latest application trends:
• P2P, Web 2.0, YouTube, etc.
• We focus on underlying infrastructure (IP, etc.)
• Building a network that can support the next
big thing without knowing what it is in advance
• Our record of research impact is mixed
Recent Impact of Networking Research
• Better technologies: optical, wireless, switches
• Better understanding: measurement, theory
• Incrementally deployable: AFD, TFRC, IDS
• Radical architectural innovation: none
Why have we failed?
Failure of Need?
• Many think the current Internet architecture will
never be sufficiently secure or reliable
• Must also deal with many new requirements:
• Ubiquitous embedded sensors
• Optical switching
• New wireless technologies
• ……..
Failure of Imagination?
• Many promising ideas in the literature
• New ideas being floated in every conference
• Some are “clean-slate” redesigns of the Internet
Failure of Deployment?
• Not really…..
• Even if we were put in charge of the Internet, we
wouldn’t know what new architecture to adopt!
Failure of Evaluation!
• Conferences are littered with promising proposals
• But we can’t tell the good ideas from the bad
• Because we never see them in operation
• Evaluations: simulation and toy-deployments
• Architecture is no longer an experimental science
• It has become science fiction
The Research Cycle
A Piece is Missing!
Deployment
Design
Simulation / Emulation
(results)
(measurements)
(code)
Experiment At Scale
What About Traditional Testbeds?
• Production testbeds:
• Can’t try radical experiments
• Experimental testbeds:
• No real users
• Not much better than simulation
• Both kinds of testbeds:
• Only one experiment at a time
• Limited to sites directly connected to testbed
• Hard to program
What We’d Really Like
• Usable by many experiments simultaneously
• Easily programmable
• Can experiment on any level (optical to apps)
• Users can “opt-in” even from remote locations
• Reasonably large scale
Be Careful What You Ask For…..
• NSF is proposing to build a large experimental
facility for networking research
• GENI: Global Environment for Network Innovations
• Funded by NSF’s Major Research Equipment and
Facilities Construction (MREFC) account
• This would be computer science’s first MREFC
GENI Design Principles
• GENI is comprised of network resources
• Links, nodes, subnets,…
• Resources are virtualizable and programmable
• Can be partitioned among many researchers
• Can implement radical new designs
• Researchers can program GENI at any level of
abstraction
• Optical, IP, application,….
GENI Design Principles (cont’d)
• Wide variety of networking technologies
• Optical, wireless, sensors, phones,…
• Large-scale (~25 PoPs)
• Users can access GENI through overlay
This is PlanetLab on Steroids!
Each Researcher Gets a “Slice” of GENI
And They Don’t Interfere
User Opt-in
Clien
t Proxy
Serve
r
National Fiber Facility
+ Programmable Routers
+ Clusters at Edge Sites
+ Wireless Subnets
+ ISP Peers
MAE-West
MAE-East
GENI Will Enable Us To…
• Experiment at scale
• 1000s of simultaneous experiments
• Long-running services (operational experience)
• Integrate our designs
Process
• GENI still undergoing evaluation and review
• Soonest funding in 2009
• Relevant bodies:
• Interim planning group
• GENI Science Council
• GENI Project Office (bid not awarded yet)
• What about the research on GENI?
NSF Funding Architectural Research
• NSF’s Future InterNet Design (FIND) Program
• Funding “clean-slate” research
• No constraints on backwards compatibility
• Effort being coordinated by David Clark
Example Ideas
• Accountable and diagnosable Internet
• Data-oriented pub/sub Internet
• Internet without addresses (only names)
• Secure enterprise networks
• Super-robust routing algorithms
Clean-Slate as Means, Not End
• No one expects direct adoption of radical ideas
• It is the ideas that will have impact
Clean-slate designs  Insights  Better Internet
Herding Cats
• Facility and funding are not enough
• Community must work together
• Architecture is not simple sum of 300 papers
• Product of synthesis and collaboration
• FIND’s goal is to have community eventually
develop a “few” coherent architectures
What Does This Mean for Us?
Two important changes:
• From science fiction to engineering
• More experimental focus, enabled by GENI
• From publish-or-perish PIs to community effort
• Drive towards coherent architecture
This is the future of Internet reseach!
What Does This Mean for You?
• A better Internet….
• ….we just don’t know what it looks like yet.
For More Information
• www.geni.net
• Discussion Mailing List
• GENI Design Document (GDD) Series