NSF Workshop on Future Directions in Network Modeling, Simulation and Measurement
August 15, 16, 2005
NSF Headquarters, Arlington Va.
Room 595

Overview.

The workshop will include approximately 25 persons from academia, government, and industry.  It will be held at the National Science Foundation Headquarters in Arlington Va. (4201 Wilson Blvd.) on Monday and Tuesday, August 15 and 16, 2005,with final wrap up and conclusions no later than 1:00pm on August 16.The workshop is in room 595.

Goals.

In recent years DARPA and NSF funded programs have resulted in substantial advances in our ability to model and simulate large-scale computer communication networks.  Such advances  have the potential to lead to fundamental advances in our  understanding of the behavior of such systems, thereby offering new opportunities to create more reliable, robust, and secure networks in the future.

The goal of this workshop is to bring together developers and potential users of advanced network modeling and simulation technology in order to both assess the current state of the art and to provide recommendations concerning the next steps that should be taken to maximize the impact of such technologies within the networking research community and industry. For example, one possible future direction might be the establishment of one or more testbeds providing large-scale network simulation capabilities to users.  One specific area of particular interest concerns the application of network modeling to the design of new Internet architectures.  Another specific area concerns the application of network modeling and simulation in information assurance applications.

Structure of the Workshop.

The workshop will consist primarily of several small group "breakout" sessions, with specific objectives for each one, identified below.  Prior to the breakouts, there will be four short presentations each discussing the current state-of-the-art in a specific area, and some background useful for the breakout sessions.  After the presentations, there will be a group discussion of all participants identifying specific items for each breakout group to discuss and analyze.  On day 2 of the workshop, there will be two short presentations by government representatives,  giving an overview of the modeling and simulation needs of their respective organizations.  These will be followed by two more breakout sessions, and finally a wrapup session with short reports from breakout chairs and conclusions.  The data product from the workshop will be a detailed report to NSF outlining a research agenda in the modeling and simulation area for consideration for future NSF programs.

Breakout Sessions.

Session 1, Technology Development.

The focus of this breakout is the discussion of research directions for network simulation tools and their transition to the broader networking research community.  What are the primary design goals for future tools?  What should be the focus of future research in simulation tools; should it be on high-fidelity models?  low-fidelity models for faster performance; mixed models?  Is performance the primary issue?  Is "the bigger the better" in terms of scale? Is extensibility and usability a primary concern?  What would it take to create a new, scalable, extensible tool for network simulation that would meet the needs of the overall networking research community?

Session 2, Emerging Applications.

This group will focus on newer and increasingly important applications, and how modeling and simulation can be used in those areas.  As an example, the security and survivability of the Internet is becoming increasingly important, as is a "new design" for the Internet.  Can existing modeling and simulation tools be used effectively in these (and similar) research areas?  What is needed that is not presently availble? How can researchers in the security community (for example) communicate effectively with the simulation technology developers to insure we have useful and effective tools for research in that area. Other emerging "hot" application areas should be identified, and evaluations of simulation capabilities to meet the needs of those areas should be discussed.

Session 3, Measurement and Validation.

This group should focus on steps needed to insure that the models accurately reflect the performance and behavior of deployed networks in similar conditions.  How can existing measurement research efforts be leveraged to assist with validation.  How can measurement datasets be used to create more realistic user behavior models?  Do we need detailed performance models of a given Cisco router model (for example) to insure accuracy in the simulations?  How can we insure that the performance of the TCP stack in a given Windows implementation (for example) is accurately reflected in the TCP model for a simulation tools. Is this level of verification necessary?  How do we insure the feature enhancement and bug fixes in simulation tools do not introduce inconsistencies or errors in results?

Session 4, Testbeds.

Testbed and emulation environments are important tools that are widely used within the networking research community. Are existing testbeds meeting the needs of the community?  What needs to be done to enhance these tools to make them more useful and more generally available? Can testbeds play a roll in any future "redesign" of the Internet?

Agenda.

Monday, Aug 15

 

 

8:30-8:45

Breakfast

 

8:45-9:00

Opening Remarks

Darleen Fisher (NSF)

Don Towsley (UMass), Richard Fujimoto (GT), George Riley (GT)

9:00:10:30

Background Presentations

 

 

Wireless Simulation

Rajive Bagrodia (UCLA)

 

Wired Simulation

David Nicol (UIUC)

 

Security

Douglas Maughan (DHS)

10:30-10:45

Break

 

10:45-12:15

Presentations (cont.)

 

 

Network Models

Albert Greenberg (ATT)

 

Needs for Internet Research

Sally Floyd (ACIR)

 

Notes on M&S of Networks

Jean Walrand (UCB)

12:15-1:30

Lunch

 

1:30-4:30

Parallel Breakout Sessions

 

 

Technology Development

Sally Floyd (ACIR)

 

Emerging Applications

John Baras (Univ. of Maryland)

4:30-5:30

Breakout Reports

 

Tuesday, Aug 16

 

 

8:15-8:30

Breakfast

 

8:30-9:00

Government Perspective

 

 

NSF

Wei Zhao (NSF)

9:00-11:30

Parallel Breakout Sessions

 

 

Model Validation

KC Claffy (UCSD/CAIDA)

 

Testbeds

Bharat Doshi (UMass)

11:30-1:00

Breakout Reports

Wrapup

 

 

 

Logistics.

Lometa Mitchell (lometa at cc.gatech.edu) is handling logistics for the meeting.  Original receipts should be sent to:

Lometa Mitchell

College of Computing

801 Atlantic Drive

Georgia Institute of Technology

Atlanta, GA 30332-0280

 

We recommend one of the hotels listed below that are close to the meeting location.
***************************************
Seidman Center
1001 N. Monroe Street
Arlington, VA 22201
703-516-4630
$125.00 per night
Please note that the facility is part of the FDIC government building, and all guests must have a valid U.S.
Government-issued ID in order to enter the premises.
***************************************
Holiday Inn
Arlington @ Ballston
4610 N. Fairfax Dr.
Arlington, VA 22203
703-243-9800
No GovŐt rate available: Hotel rate is $179.00 per night
***************************************
Comfort Inn Ballston
1211 N. Glebe Rd.
Arlington, VA 22201
703-247-3399
GovŐt rate: $153.00 per night
****************************************

For further information, contact one of the workshop organizers:

Richard Fujimoto (fujimoto at cc.gatech.edu)

George Riley (riley at ece.gatech.edu)

Don Towsley (towsley at newworld.cs.umass.edu)