Transcript Clean Slate Design for the Internet

Clean Slate Seminar
CS541: Fall 2007/8
Nick McKeown
[email protected]
Guru Parulkar
[email protected]
http://cleanslate.stanford.edu
Outline
 What
is clean-slate research?
 Stanford’s Clean Slate Program
 Logistics
for CS541
– Schedule for the quarter
– What we expect from you
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Anything to rethink?
Well-known researcher in mid-1990s:
“Multimedia communications over the Internet is done.”
“How come it takes an hour to set up a session?”
“Why can I join someone else’s call?”
“Will the quality always be this poor?”
“Can I put a camera on my car and drive around?”
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Story
The Internet enabled universal communications
and transformed society
 Other infrastructure technologies have had big
impact (e.g. electrification, water supply and
distribution, highways/cars, radio & TV,
telephone)
 All have changed enormously over time …
except the Internet
 The Internet infrastructure will change
 The only question is how.

CS541: Clean Slate Research
http://cleanslate.stanford.edu
The Internet Hourglass
Applications
Kazaa VoIP Mail
TCP
News
Video
Audio
IM
YouTube
Everything
Transport protocols
on IP
SCTP
UDP
ICMP
IP
Ethernet 802.11
IP on
Power lines ATM Optical
everything
Satellite Bluetooth
Link technologies
CS541: Clean Slate Research
Modified John Doyle Slide
http://cleanslate.stanford.edu
Internet and WEB Hourglass
Kazaa VoIP Mail
Applications
Everything
Newson Video
WEB Audio
IM
YouTube
HTTP
Little change
Transport protocols
TCP
SIP
UDP
RTP
Continued
Innovations
IP
Ethernet 802.11
on
Power lines IP
ATM
CS541: Clean Slate Research
everything
Optical
Satellite Bluetooth
Modified John Doyle Slide
http://cleanslate.stanford.edu
Internet Technology Innovations

Optical transmission and WDM

Packet switching

Wireless and mobile networking

Internetworking architecture

Routers, switches, security devices, routing systems

Domain name system

Client/server architectures and applications

HTTP, browsers, servers

Search engines, targeted advertisements

P2P technologies and applications

And many more
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Internet Architecture Limitations

Security & robustness - to support other critical infrastructures

Control and management

Addressing, naming & (inter-domain) routing

End-to-end principle vs in-network processing

Mobility of hosts and networks

Economic viability of different stakeholders
CS541: Clean Slate Research
http://cleanslate.stanford.edu
State of Internet
“… in the thirty-odd years since its invention, new uses and
abuses, …, are pushing the Internet into realms that its original
design neither anticipated nor easily accommodates.”
“Freezing forevermore the current architecture would be bad
enough, but in fact the situation is deteriorating. These
architectural barnacles—unsightly outcroppings that have
affixed themselves to an unmoving architecture— may serve a
valuable short-term purpose, but significantly impair the longterm flexibility, reliability, security, and manageability of the
Internet.”
Overcoming Barriers to Disruptive Innovation in Networking, NSF Workshp Report, 05.
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Unthought of applications
Economically sustainable
Trustworthy: Secure, robust, manageable
Mobility by default. Users and data
Performance to blow our socks off
Unthought of links
CS541: Clean Slate Research
http://cleanslate.stanford.edu
A Research Problem, or
A Problem with Research?
Cultural problem
– Research community was stuck in incrementalism
and backward compatibility
– Researchers need to ask hard and unpopular
questions
– Compatibility with reputation and career prospects…?
Yet, radical approaches are common in other fields
Is it just a problem of installed base?
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Rethinking the car
Installed base
1 gallon of gas g 22lbs of CO2
1900
1968
2007
8,000
170M
700M
Car
Engine
Control
Fuel
Car Body
Materials Manufacture Safety
CS541: Clean Slate Research
Policy
Emissions
Fueling
Stations
http://cleanslate.stanford.edu
General Approach in Research
Community
1.
Build a nationwide research facility upon
which new architectures can be tested
– NSF GENI Program http://www.geni.net
– Key concept: Virtualization to support
multiple architectures simultaneously.
2.
Deploy and test experimental
architectures
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Conception-to-Deployment
Case for GENI Facility
Maturity
Need for Large experimental
facility/infrastructure
Shared
Deployed
Infrastructure
This chasm represents a major
barrier to impact real world
Small Scale
Testbeds
Research
Prototypes
Foundational
Research
Time
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Facility Design: Key Concepts
Sensor Network
Edge Site
Mobile Wireless Network
Federated Facilities
Slicing, Virtualization, Programmability
CS541: Clean Slate Research
http://cleanslate.stanford.edu
The Stanford Clean Slate Program
Approach
Motivating Questions
“How will the Internet look in 15-20 years?”
“With what we know today, if we started with a clean-slate, how would we
design a new Internet?”
Bring together Stanford’s breadth and depth:
Networking, optical communications, wireless, access networks, theory,
economics, security, applications, multimedia, operating systems, hardware
and VLSI, system architecture, …
Research for long term impact on the practice of
networking
Two pronged approach: “innovations in the small” and “innovations in the large”
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Broad Interdisciplinary Focus
Heterogeneous
Applications
Security
Robustness
Network
Architectures
Economics
Policies
Heterogeneous
PHY Technologies
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Clean Slate Program Approach
“Innovation in the small”
Architectural
 Start many small
exploratory projects
Blueprint?
 Lots of new ideas
 Lots of areas
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Clean Slate Approach Cont
“Innovation in the large”


Architectural
Start a few select larger
collaborative projects
Teams of students, faculty
and industry
Blueprint?
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Clean Slate Approach
Security
Backbone
(Lightflow)
Wireless
Network
Virtualization
Flow Control/Security Cheap
Backbone Congestion Wireless
Theory
(Ethane)
Prog H/W
(VLB)
Control (Spectrum)
(NetFPGA)
(RCP)
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Example projects
1.
2.
3.
4.
“If we started over, how could we design
a more secure network?”
“How could we design a network good
enough for tele-surgery?”
“What is the economic/business model
that will lead to multiple, concurrent
virtualized networks?”
“What platforms facilitate others to do
clean-slate research?”
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Example Project
Facilitating others
Programmable building blocks
 XORP: Software, extensible, open-source
 WARP: Programmable wireless MACs
 WUSTL Programmable router
 NetFPGA: Programmable network hardware
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Outline
 What
is clean-slate research?
 Stanford’s Clean Slate Program
 Logistics
for CS541
– Schedule for the quarter
– What we expect from you
CS541: Clean Slate Research
http://cleanslate.stanford.edu
CS541: Schedule for Quarter
Week 2 (Oct 2) “GENI Research Rationale”
Ellen Zegura, Georgia Tech, Co-Chair GENI Science Council
Week 3 (Oct 9) “GENI Research Rationale for Reinventing the Internet”
Stanford Faculty Panel
Week 4 (Oct 16) TBA
Week 5 (Oct 23) James Kempf (DoCoMo Labs) + Student Panel
Week 6 (Oct 30) “Mobile Wireless and Its Implications on Future”
D. Raychaudhuri, Rutgers, Co-Chair of GENI Mobile Wireless Working Group
Week 7 (Nov 6) “Data Center Architectures”
Alan Berstein, Credit Suisse
Week 8 (Nov 13) TBA
Week 9 (Nov 20) “Distributed Systems and Future Internet”
Frans Kaashoek, MIT
Week 10 (Nov 27 or Dec 4) “Distributed Services for GENI”
Amin Vahdat, University of California San Diego.
CS541: Clean Slate Research
http://cleanslate.stanford.edu
Format for talks and panels
4 – 4.15pm: Refreshments
4.15 – 5.15pm: Guest speaker
5.15 - 6pm: Reverse Panel
Student panel questions our guest speaker
CS541: Clean Slate Research
http://cleanslate.stanford.edu
What we expect from you
If you are taking for credit…
 Attend every week (you are allowed to
miss one)
 Write a 5-page report on a clean-slate
topic, agreed with instructors, or
 Take part in student panel
CS541: Clean Slate Research
http://cleanslate.stanford.edu