5/17/2001NLANR & I2 Joint Techs Bandwidth ahoy! …NOT!! Ana Preston The University of Tennessee May...
-
Upload
martin-jones -
Category
Documents
-
view
213 -
download
0
Transcript of 5/17/2001NLANR & I2 Joint Techs Bandwidth ahoy! …NOT!! Ana Preston The University of Tennessee May...
5/17/2001 NLANR & I2 Joint Techs
Bandwidth ahoy! …NOT!!
Ana PrestonThe University of Tennessee
May 17, 2001
5/17/2001 NLANR & I2 Joint Techs
At Univ. Tennessee Early days: tried blocking on well-
known Napster ports and well-known IP addresses
For dorm traffic: applied CAR per subnet: 20Mbps (before Resnet-DS-3)
Most efficient method: DoS attack on Napster ;-)
5/17/2001 NLANR & I2 Joint Techs
At Univ. Tennessee, cont. Homegrown flow-based graphs SNMP queries on all campus routers
Wed. Feb. 14, 2001 - Qwest-DS-3
Fri. March 2, 2001 - Qwest-DS-3
Fri. March 2, 2001 - Resnet-DS-3
Fri. April 27 - Qwest-DS-3
Fri. April 27 - ResNet-DS-3
Thurs, May 10, 2001 - Qwest-DS-3
Thurs., May 10, 2001 - ResNet-DS-3
Meanwhile, Internet2 DS-3, Feb. 14, 2001
Meanwhile, Internet2 DS-3, Frid. April 27, 2001
Meanwhile, Internet2 DS-3, Thurs. May 10, 2001
5/17/2001 NLANR & I2 Joint Techs
Univ. of Tennessee - Fri. March 2, 2001
5/17/2001 NLANR & I2 Joint Techs
Today: P2P is very much alive... How universities have responded:
Banning it (mainly b/c it clogs the network) “Passively” Monitoring (and keeping tabs on
top hogs…) Blocking, Rate Limiting, Fair Usage Policy
Implementations Educating (on copyright law and impact.) Not doing anything at all and waiting…
5/17/2001 NLANR & I2 Joint Techs
The challenges... Should you control p2p applications and if so, on
what basis? Is this the right question anymore? planning for constantly increasing bandwidth
demand without increasing funding. Bandwidth: for students (and staff and
faculty/reserarchers): How much? What is fair? How to implement?
P2P applications are changing our traditional paradigms.
[archives at http://listserv.utk.edu/archives/p2p.html]
5/17/2001 NLANR & I2 Joint Techs
Nothing new, yet... advent of P2P applications is challenging some of the
broader Internet architecture models
“Universities have been the first to face these challenges, and as a result, universities are increasingly looking [or could be looking] into new and innovative architecture and service models encompassing notions like settlement free exchange-point based peering, massive peering, GigaPOPs, dark fiber nets, neutral collocation facilities, self organizing nets, etc.” (from a post to the p2p list)
5/17/2001 NLANR & I2 Joint Techs
Mark your calendars Oct. 4-5, 2001, following the Fall Internet2 Member
meeting at Austin, TX (Univ. of Texas at Austin, Sept. 30 - Oct. 4, 2001).
The first R&E p2p workshop!
5/17/2001 NLANR & I2 Joint Techs
The p2p workshop! Our goal:
We will specifically explore the technical and future dimensions of the fast-growing P2P services spaces and the opportunities and challenges presented for universities.
5/17/2001 NLANR & I2 Joint Techs
What this will be about P2P application developers, users and us (??) in the
same room to share experiences and challenges. Exploration into how universities can integrate and
perhaps even innovate with this "revolution,” while providing vehicle (or models, ideas, and such) through which universities can benefit.
Advantages of P2P models for applications in research, learning and teaching.
What could there be for the R&E community for future opportunities?
5/17/2001 NLANR & I2 Joint Techs
A. How do these things work? In general, how are P2P applications designed and what
are the basic kinds of architectures/models that are used.
What are the issues and challenges that come up when creating P2P systems? Are there ways that standardized architectures can be promoted that allow for collaboration between universities and P2P application developers?
Network engineers have vested interest in understanding how P2P technology operates so that they can make networks operate efficiently, perhaps in spite of P2P applications being used.
5/17/2001 NLANR & I2 Joint Techs
B. P2P in the real world How are the most innovative models are being used "in
the real world"? What are the policy and implications when these
applications are used in real case scenarios. Research and science (e.g., genome@home, climate simulation,
economics, medicine) Academia, education and learning - K-20 and more: educational
p2p applications --LEARNSTER (see www.educommons.org)- what are the advantages of P2P systems for learning and educating.
Enterprise: Groove, Porivo, OpenCola, Parabon, etc. Other...
5/17/2001 NLANR & I2 Joint Techs
C. Approaching use and innotivation
To provide a survey of successful methods that are being used at universities
The issues surrounding the ever-growing demand for bandwidth
What does it take to maximize both commodity and Internet2 bandwidth?
With P2P, where will this innovation take us and how do we take advantage of the possibilities; how do we get people to think of P2P in innovative ways? who should we be getting interested in that innovation?
5/17/2001 NLANR & I2 Joint Techs
Program Committee Chairs:Ana Preston, University of TN Linda Roos, OARnet
George Brett, NLANR-DAST Perry Brunelli, Univ. Wisconsin-Madison James Deaton, ONEnet David Futey, Kent State University Doyle Friskney, Univ. Kentucky William Green, Univ. Texa Chris Rapier, Pittsburg Supercomputing Center/NLANR Joe St Sauver, Univ. Oregon Jerry Sobieski, Univ. Maryland Steve Wallace, Indiana Univ. David Wiley, Utah State Univ
5/17/2001 NLANR & I2 Joint Techs
Interested in more? Thurs. in-depth session 1:30 p.m
Case studies P2P apps. on I1 and I2 (and implications, e.g. SEGP and p2p) Feedback on p2p FAQ and workshop
P2P list:listserv.utk.edu/p2p/archives.html
Email me at [email protected] or Linda Roos at [email protected]
Stay posted for our announcement
Thank you!