Experiences with Internet and Intranet Mapping
Bill Cheswick
Lumeta Corp

Abstract:

The Internet Mapping Project started at Bell Labs in 1997 to collect long term Internet topological data and visualize the results. In 1999, we focused on the Serbian internet during the NATO bombing, and observed major connectivity disruptions. In 2000, this technology was spun off to a startup company, Lumeta. Since then, Lumeta has enhanced this technology and scanned over fifty large corporate intranets, and a number of critical government networks. We can now report generalized characteristics of these large networks, showing variations in network management techniques and control over large networks. We are now working on tools to configure the honeyd anti-hacking tool to emulate these large networks.

Biography:

Ches logged into his first computer in the fall on 1968, and was graduated from Lehigh University in 1975. In 1987 he joined Bell Labs, where he worked on early firewall and IDS designs. With Steve Bellovin, he wrote Firewalls and Internet Security: Repelling the Wily Hacker, a fairly popular book on the subject. The second edition came out in March 2003 with the help of a third co-author, Avi Rubin. Ches also worked on commercial munitions, application gateways, PC viruses, and a variety of other Internet diversions. He started the Internet Mapping Project with Hal Burch in 1998. In 2000, he co-founded Lumeta, a spin-off from Bell Labs to commercialize various intranet exploration techniques. In his spare time, Ches flies RC airplanes, computerizes his house, and has scanned in over 80 GB of images of old photos. He is partial to steak (medium rare), and mint imperials.

Towards building a secure Internet.
Alan Boulanger
IBM Research

Abstract:

The Governments and businesses racing to use the Internet have chosen a dangerous track. The current Internet is an inherently insecure and unstable platform on which to conduct business critical operations. There are several bridging technologies that have been developed to mitigate the vulnerability of current security exposures, however large security gaps remain. These gaps can only be addressed though significant changes in the underlying architectures of the systems that comprise the Internet. This need for security is now a powerful force influencing current and future technical innovation. Many companies deploying new applications and systems are beginning to build security into the product in the design phase. Organizations are beginning to understand that security is important. The recent news reports of high profile security related incidents, such as the Melissa Virus and DDOS attacks, has brought the issue of computer security into public view. Once restricted to the domain of fantasy books and fiction, hackers and virus authors have become an increasingly visible threat to everyday users. Why is this possible? How can we as a community protect ourselves? This talk will address the current security related problems with the current implementation of the Internet and the technologies that are currently under development, along with future technologies, that are designed to make the Global Internet an safer place for users and businesses alike.

Biography:

Alan Boulanger joined IBM in October 1995 as a research member of the TJ Watson Global Security Analysis Laboratory. His research interests include network security, intrusion detection systems, applied penetration testing tools and techniques, data forensics, telephony related security, and researching new system vulnerabilities. As a result of his research, Mr. Boulanger has numerous filed patent applications related to computer security issues. Since joining IBM, Mr Boulanger has provided technical assistance to numerous Federal Law Enforcement and Intelligence Agencies and Businesses conducting computer security related investigations. As a result of his efforts, Mr. Boulanger has recieved many awards and commendations from IBM and Government Agencies. Mr. Boulanger is an invited long standing member of the New York Electronic Crimes Task Force.

Network Security: More than an End-to-End Problem
Tom Tarman
Sandia National Labs

Abstract:

Network security is often regarded as an "end to end problem," meaning that if the endpoints perform appropriate cryptographic and key management, all network security problems are solved. However, network security involves more than encrypted pipes and end-to-end authentication - it requires protection of the network infrastructure as well. This talk describes the shortcomings of regarding network security purely as an end-to-end problem, presents mechanisms for protecting network infrastructure, and introduces future research challenges in securing network infrastructure.

Biography:

Tom Tarman is a distinguished member of the technical staff at Sandia National Laboratories, in Albuquerque New Mexico, where he primarily performs network security research. Tom has been active in the field of ATM network security for the past eight years, having published several conference papers and journal articles on topics such as high-speed ATM encryption, algorithm-agile ATM encryption, and ATM security protocols. In addition, Tom has been an active participant in the ATM Forum Security Working Group, where he has authored numerous contributions and has served as editor for the ATM Security Specifications Versions 1.0 and 1.1. Tom received the ATM Forum's Spotlight Award for his technical and "PR" contributions to ATM security, and has recently co-authored (with Edward Witzke) a book entitled "Implementing Security for ATM Networks," available from Artech House Publishers. Tom's current research interests include security for MPLS and all-optical networks, network modeling and simulation, and networked multimedia applications.

Simulation of Ultra Large Networks: Simulation, Emulation and Modeling
Erol Gelenbe
University of Central Florida

Abstract:

We will discuss some of our ideas for testing advanced network techniques in "the large" by combining simulation and emulation techniques which allow novel non-IP protocols to be experimented in a realistic setting combining the Internet with experimental systems. Some related experiments that we have conducted on Cognitive Packet Networks will be presented. If time permits, we will also discuss some new research directions in the theory of network QoS.

Biography:

Erol Gelenbe (FIEEE, FACM) has served as the Nello L. Teer Professor and Chair of Electrical Engineering at Duke University (1993-98) and as the University Chair Professor of EECS and Founding Director of the School of EECS at UCF (1998-2003). His research interests include self-adaptive and autonomic networks and systems, as well as performance modeling and simulation.

Secure Networked Systems and the Future
Sumit Ghosh
Stevens Institute of Technology

Abstract:

Networked Systems are here to stay, not for the next 50 or 100 years, but for thousands of years into the foreseeable future. They will be an integral part of us and it is critical that we design it correctly. The origin of many of the winding and confusing roads in New Jersey may be traced back to the colonial days when the world was a very different place and most people did not have a clue of what U.S. was going to become. Over the past 50 years and continuing well into the future, drivers along these roads will be confused, delayed, lost, and angry. The cumulative cost is unimaginably high, a very heavy burden for the design decisions of the past. Networked systems are literally the road systems of tomorrow and, it is imperative, that we make every effort not to impose the slightest burden on the future. This presentation will focus on the fundamental attributes of secure networked systems, the challenges that arise from these attributes, and new strategies to deal with them.

Biography:

Sumit is the Thomas E. Hattrick Professor of Information Systems Engineering at Stevens Institute of Technology. He is the author of "Principles of Secure Network Systems Design" (Springer-Verlag, April 2002), "Modeling and Asynchronous Distributed Simulation" (IEEE Press, June 2000), "Algorithms for Networked Information Technology Systems" (Springer-Verlag, Aug 2003), "Intelligent Transportation Systems: New Principles and Architectures" (CRC Press, Jan 2000), and "Hardware Description Languages: Concepts and Principles" (IEEE Press, September 1999). He is coeditor of "Guarding Your Business: An Architecture for Security" (Kluwer Publisher, August 2003). His research interests include network security, networking, hardware design languages, computational intelligence, engineering creativity, ethics, and engineering education.

Embedded Values: the importance of a multi-discipline approach to network security.
Elliot Turrini
MDM

Abstract:

Our current and future reliance on digital networking technology has made network security an important social, legal, and economic issue. Businesses, governments, and individuals have raced to adopt digital networking technology -- with little concern for the downside. The upside of digital networking is clear: substantial improvements in communication capabilities. Unfortunately, however, the downside has been hidden, neglected, or a combination of the two. My presentation contends that a multi-discipline approach involving law, technology, psychology, economics, and risk management/insurance is required to (a) reveal the downside of digital networking technology and (b) improve the probability that digital networking technology will provide a net social gain. Moreover, I will contend that the scientists, engineers, and technologists working on digital network technology must be aware of the multi-discipline issues, so that they can incorporate that knowledge into their work. Digital networking technology, standing alone, may provide one of the most effective ways to reduce the downside of this technology. Doing so, however, requires a broad scope of knowledge, which can be applied through a multi-discipline approach.

Biography:

ELLIOT TURRINI received a Bachelor's degree from Yale University in 1987 and his Juris Doctorate Summa Cum Laude from Seton Hall University School of Law in 1992, where he was an Articles Editor for the Law Review. He is the former law clerk to the Honorable Morton I. Greenberg, United States Court of Appeals for the Third Circuit, and to the Honorable Kenneth C. MacKenzie, Presiding Judge, Chancery Division, Morris and Sussex Counties. He was previously associated with the firm of Lowenstein Sandler in Roseland, New Jersey, before joining the United States Attorney's Office in Newark where he served for seven years. During his tenure as a federal prosecutor, he prosecuted some of the Department of Justice's most significant chemical diversion cases, as well as one of the Department's largest international money laundering investigations. Moreover, he conducted complex federal criminal trials. His major accomplishments at the United States Attorney's Office came in his capacity as a Computer and Telecommunication Coordinator, where he prosecuted computer criminals including David Smith for having disseminated the Melissa Virus. As a result of his expertise in computer crime and information security, Wadsworth Publishing asked him to edit an information security/computer crime book that is due to be published summer 2003. The book is a multi-discipline contributed reader that brings together experts in law, technology, psychology, economics, risk management, and insurance to assist the private and public sectors develop efficient, effective, and responsible computer crime/information security strategies. At MDM, as more fully described below, Mr. Turrini's practice will focus on two areas: (1) Information Security, Cyberlaw, and Privacy, and (2) Controlled Substances -- such as pseudoephedrine and ephedrine. He will also be doing corporate investigations, complex civil litigation, white collar criminal defense, and anti-money laundering compliance/counseling.

Large-Scale Network Simulation: How Big? How Fast?
Richard Fujimoto
Georgia Institute of Technology

Abstract:

Parallel and distributed network simulation tools are emerging that offer the ability to simulate networks containing millions of network nodes and hundreds of thousands of concurrent traffic flows in real- or near-real-time. This capability offers enormous opportunities for researchers to study scalability issues that could not be previously addressed. At the same time, it also creates challenges to the networking research community to create scenarios and configurations that are realistic relative to current and future Internet configurations. It creates challenges to tool builders to create verified and validated simulators that are easy to use and execute efficiently on parallel and distributed computers over a wide range of network configurations and scenarios. This presentation will describe an approach to realizing scalable network simulations that leverage existing sequential simulation models and software. Specifically, two parallel network simulators have been developed, one based on the widely used ns2 simulator (termed pdns), and another based on a tool developed at Georgia Tech called GTNets. Packet-level simulations using pdns executing on 1024 processors at the Pittsburgh Supercomputer Center yielded performance as high as 80 Million simulated packet transmissions per second of wallclock time for a network containing over 3.8 million network nodes. This research represents joint work with Drs. Mostafa Ammar, Kalyan Perumalla, George Riley and several PhD students at Georgia Tech, and is funded by NSF (grants ANI-9977544 and ANI-0136939) and DARPA (contract N66001-00-1-8934).

Biography:

Dr. Richard Fujimoto is a professor in the College of Computing at the Georgia Institute of Technology. He received the Ph.D. and M.S. degrees from the University of California at Berkeley in 1980 and 1983 in Computer Science and Electrical Engineering, and B.S. degrees from the University of Illinois at Urbana in 1977 and 1978 in Computer Science and Computer Engineering, respectively. He has been an active researcher in the parallel and distributed simulation community since 1985, and has published numerous technical papers as well as a book on this subject. He has led the development of parallel/distributed simulation software systems including the Georgia Tech Time Warp (GTW) simulation executive on which the TeD parallel network simulator is based, and the Federated Simulation Development Kit (FDK) used to create parallel versions of ns2 and GTNets. He has given several tutorials on parallel and distributed simulation at leading conferences. He led the definition of the time management services for the U.S. Department of Defense High Level Architecture (HLA). Fujimoto is Co-Editor-in-Chief of SCS Transactions (as of July 1, 2003), and has been an area editor for ACM Transactions on Modeling and Computer Simulation since it was founded in 1990. He has served on the organizing and program committees of several major simulation conferences such as the Workshop on Parallel and Distributed Simulation (PADS) and the Simulation Interoperability Workshop (SIW).

Modeling and simulation of security in ULN: can we apply ideas of how brains focus attention on the most urgent inputs?
Bernard Zeigler
University of Arizona

Abstract:

Concepts in cognitive science have been developed on how the brain focuses its perceptual resources on the most active elements of its sensory inputs. Such mechanisms have been shown to explain how visual search, otherwise computationally intractable, is rendered feasible. Taking the analogy between the brain and a large scale network one step further, we have developed some distributed attention management mechanisms and studied them via modeling and simulation. The results suggest how detection of threats in ULNs can be implemented in a distributed, scalable manner.

Biography:

Bernard P. Zeigler is Professor of Electrical and Computer Engineering at the University of Arizona, Tucson and Director of the Arizona Center for Integrative Modeling and Simulation. He is internationally known for his 1976 foundational text Theory of Modeling and Simulation, recently revised for a second edition (Academic Press, 2000), He has published numerous books and research publications on the Discrete Event System Specification (DEVS) formalism. In 1995, he was named Fellow of the IEEE in recognition of his contributions to the theory of discrete event simulation. In 2000 he received the McLeod Founder's Award by the Society for Computer Simulation, its highest recognition, for his contributions to discrete event simulation. In June 2002, he was elected President of the Society (recently, renamed The Society for Modeling and Simulation, International.) In 2003, his autobiographical retrospective and the evolution of the theory of modeling and simulation will appear in the International Journal of General Systems.

Changes Required to Secure the Distributed Intelligent Network of the Future
Bill Way
Bytex Corporation

Abstract:

The net.CARE system provides the critical information needed to control and manage today's complex & dynamic high-performance enterprise networks. I will present how this system has evolved from work done at NSA, DISA, and Sandia. The issues I would note are the different owners (stakeholders) of each network layer and intermediate transports. Thus potentially conflicting security goals. I will also address our use of models and simulation (micro vs. macro modeling). The fact that most simulations do not perform well in the end-case situations of the real networks. Those being congestion, buffer overflows, interfaces that are not clearly defined. The simulation tool costs are such that they are generally used only by vendors and are representative of only that vendors' products. Issues faced in providing security for the future optical networks such as GigBE. This is one DOD version of an Ultra Large Network.

Biography:

Bill Way has forty years of work experience in computers, supercomputers and networking. Served in the role of Business Development for several high tech companies and enjoyed dialog with almost every computer and networking company in the capacity of buying , licensing or selling technology or potential M&A. During the last twenty years I have been involved in product planning. First for Network Systems with HyperChannel and Hippi - then internetworking. At Network we were early leaders in packet filtering and provider of the first NSA network. Network Systems acquired Bytex, Vitalink and BusTech to fill out their internetworking line. StorageTech acquired Network System to create storage networks (SANs). I played an active role in this process learning from many of the founders of internetworking. I acquired Bytex three years ago with the mission of developing broadband management and security tools. The system has evolved starting with Network Systems research for NSA and the development of the first ATM Firewall then adding the NSA research on mapping and monitoring and Sandia's effort in SNIDE. He holds a Masters Degree in Econometrics from University of Minnesota.