Transcript Slide 1

Networking Research Center
Industry Day
Penn State University
College of Engineering
David N. Wormley, Dean
Tuesday, October 17, 2006
IndustryDay.ppt
NATIONAL POSITION
 PSU ENGINEERING ranks among the
largest and highest quality colleges in
the U.S.
 04/05 U.S. Engineering Degrees
Awarded
• 2nd in B.S. degrees awarded
• 7th in Ph.D. degrees awarded
 U.S. News Engineering Rankings
• 14th in U.S. undergraduate
• 19th in U.S. graduate
Undergraduate Curriculum
Prepare World-Class
Engineers:
Aware of the World
Solidly Grounded
Technically Broad
Versatile
Effective in Team
Operations
Effective in Leadership

FIRST YEAR INTERNATIONAL
DESIGN
• Collaboration with French students from
Universite D’Artois work on Industry
Sponsored Projects
• Design Reports written in both English
and French
• Collaborations with 4 other EU
Universities


INTERNATIONAL INTERNSHIPS
FIRST YEAR SEMINAR ON
ENGINEERING IN CHINA
LEARNING FACTORY



Experiential Learning
Teams from Penn
State, U of
Washington, U of
Puerto Rico, Mayaguez
developed Learning
Factory
In 2006, teams
received National
Academy of
Engineering $500,000
Bernard Gordon Award
for Educational
Innovation
Interdisciplinary Minors

Engineering Leadership Development
• Intended for students to supplement their major
field of study with knowledge of leadership
concepts, principles, practices, and techniques
• Current enrollment: 158 (13% non-engr)

Engineering Entrepreneurship
• Intended for students aspiring to be innovation
leaders of new technology-based products and
companies
 Current enrollment: 151 in core courses;
80 in higher level electives
ENGINEERING:….A Creative Profession, Serving Society
Research Thrusts in
Areas of State and National Needs

Energy/Environment

Nanotechnology

Biomedical/Biotech

Information Technology

Homeland Security
ENGINEERING:….A Creative Profession, Serving Society
Growth in Research Activities
Research in Millions
 Recent increases
represent significant
growth in
interdisciplinary
research.
 Industry funds ~ 20%
of COE research.
100
80
60
71.4
78.2
85.2
92.4 92.6
55
40
20
0
'00 '01 '02 '03 '04 '05
Penn State Engineering:
Partnering with Industry on Research

Central Industrial Research
Office
• Serves as liaison between PSU and
industries

PSU – 2nd in industry sponsored
research in nation
Penn State Engineering:
Partnering with Industry on Research



Gifts – support a variety of faculty activities
Unrestricted Grants – support research activities
Cooperative Agreements
•

Sponsor involvement on a collaborative basis
Master Agreements with IP pre-negotiation to support research
through Task Orders.
• Currently 45 Master Agreements are in place including:

Northrup Grumman

Kennemetal

Toyota Manufacturing NA

Pratt and Whitney

Lockheed Martin

Raytheon

GM

Corning, Inc.

Hershey Foods

Pittsburgh Digital Greenhouse
Network Education and
Research Programs
throughout the College of Engineering

Activities centered in:
• Aerospace Engineering
• Computer Science & Engineering
• Electrical Engineering
• Industrial & Manufacturing Engineering
• Mechanical & Nuclear Engineering
Department of
Aerospace Engineering
Aerospace applications
Fleets of small, low-cost Uninhabited
Aerial Vehicles (UAVs) for:
•Meteorology
•Surveying and remote sensing
•Surveillance and target tracking
Satellite Constellations
•Terrestrial Planet Finder
•Darwin (ESA mission)
CSE Research Areas
Research Centers/Labs

Center for Embedded and Mobile Computing
(CSE, EE, Mat Sci, ME, Civil)
• Co-Directors: Anand Sivasubramaniam and Janie Irwin

Center for Networking & Security (CSE, EE,
Statistics, IST)
• Director: T. La Porta

Center for Machine Learning Applications (CSE,
IST, Statistics, EE, Business)
• Director: R. Acharya

Scalable Computing Lab (CSE, Mat Sci, EE, Aero, Math,
Physics, Civil)
• Director: P. Raghavan

Center for Computer Vision (CSE, EE, Statistics, Civil,
ME)
Department of
Electrical Engineering
AYLIN YENER
Associate Professor of Electrical Engineering
Recent Awards (since 2005):
 Two NSF grants
 DARPA- CBMANET program
New:
 $6.5M DARPA Grand Challenge that called for a new
Information Theory for Mobile Ad Hoc Networks
(ITMANET): The winning team consists of Prof. Yener
and 11 professors from 7 other universities.
Mohsen Kavehrad
Electrical Engineering Department
Biologically Inspired Signal
Processing
Work in different areas:
Antenna Design - Doug Werner
Power Systems - Kwang Lee
Signal Processing - Ken Jenkins




Genetic optimization
algorithms
Particle swarm optimization
Ant colony optimization
Neural network pattern
recognition
Payload-Based High-Speed Network Intrusion Detection System
(IDS)
J. Wang, G. Kesidis, D.J. Miller, and I. Hamadeh
High-Speed Worm Defense by Using Both Header and Payload
Incoming
Packets
Multidimensional Traffic Clustering and Classification
Packet Header
Hashing
Multidimensional
Clustering
Suspicious Cluster Pool
Suspicious Cluster
Identification
Innocuous Packet Pool
Worm Signature Extraction and Evaluation
Worm Signature
Extration
Worm Signature
Evaluation
Worm Signatures
Payload-Based Worm Containment
Multidimensional Traffic Mining
 Frequent item set mining applied to network traffic flows, based on the
packet header 5-tuple (source IP, destination IP, source port, destination
port, protocol)
 Using top-down method to build up a multidimensional tree and only mining
significant (with traffic volume larger than a threshold) and suspicious
clusters
Pipelined Implementation
 3T delay (T could be as small as 1 second): hash packets arriving in [0, T]; perform
mining in [T, 2T]; collect suspicious packets and extract signatures in [2T, 3T]
 Requiring 50 to 60 MB memory without packet sampling
 Can handle the link with load up to 800Mbps in real-time by software
Suspicious Cluster Identification
Two Criteria for Defining a Suspicious Cluster:
 Its traffic volume is larger than a threshold (e.g., 1%)
 Its source or destination IP dispersion/cardinality is larger than a threshold.
In other words, the number of different source or destination IP addresses
involved in a cluster is larger than a threshold (e.g., 30)
Worm Signature Extraction
 Building a generalized suffix tree for each suspicious cluster, and extracting
signatures only from a small part of packets
 With time and space linear in the length of each suspicious cluster
 Jointly considering length, frequency and false positive to improve the accuracy of
signatures
 Easily modifying rules to search for multiple signature descriptions of varying length
range, with little increase in complexity
Organic Thin Film Electronics
– Electronics Anywhere
T. N. Jackson
Center for Thin Film Devices and Materials Research Institute, Electrical
Engineering, Penn State University
Low-cost devices and circuits on arbitrary substrates
Organic thin film transistor
polymeric substrate AMLCD
Electronics anywhere
OTFTs on
non-planar
surfaces
Organic display
Transistors
on cloth
a-Si:H active matrix
Gamma ray detector on
polyimide substrate
a-Si:H active matrix
OLED display
Transistors on cloth
a-Si:H strain
bridge array
Marcus Department of
Industrial and
Manufacturing Engineering
NSF-SST: PIPELINE INFRASTRUCTURE MONITORING
T-mote sensor board
Health Monitoring
6000
FT
5000
Magnitude
4000
3000
2000
1000
0
WT
Sensor data fusion and
data mining
Identify and respond
to defects
0
100
200
300
400
500
Frequency [Hz]
600
700
800
Report data
S. Kumara (PSUIE)
Xiang Zhang (UCB)
Sangkook Kim (MIT)
S. Bukkapatnam (OSU)
Sensor Network
: Clustering and
routing algorithm
Broadcasting
information
Defect
Maintenance
group
Perform maintenance
(Self-Healing)
Sensor for collecting data
Miniature Wireless Sensors (Self-sustainable)
Advanced Technology for
interoperability of Heterogeneous
Enterprise Network & their Applications
Business Object
Document Ontology
(1) Create
(2) Annotate
A*
ATHOS
(3) Def. Rules
for mapping
Local
Interface
Enterprise 1
R
(3) Def. Rules
for mapping
ARGOS
R
Local
Interface
Enterprise 2
Local
Interface
Internet
(4) Reconcile
ARES
Enterprise 1
Enterprise 2
Internet
Local
Interface
(1) Partners in USA : AIAG, NIST, GM, Ford, Microsoft, IBM, SAP, and other business partners
and software vendors (PSUIE – Kumara & Oh support GM)
(2) Partners in EU : CNR in Italy, SAP in Austria, Breza in Serbia and other laboratories in
European countries
Information Technology Related Research
Mechanical and Nuclear Engineering Department
NSF Information Technology Research Grants:
 Data-Driven Autonomic Performance Modulation for Servers, Q. Wang
 Infrastructure of the Networked Robotics & Signal Intelligence
Laboratory, A. Ray
 An Agent-Based Negotiation Framework for the Robust Design of
Active-Passive Hybrid Piezoelectric Vibration Control Networks, K-W.
Wang
Cyber-infrastructure Grant:
 Network of Design Repositories, T. Simpson
NSF-ITR: Data-Driven Autonomic
Performance Modulation for Servers
Qian Wang, MNE Department
with A. Sivasubramaniam and N. Gautam


Objective: Self-managing large
clusters of server systems such as
hosting centers & data centers
• Performance modulation
• Resource allocation
• Power management
Approach: Control-theoretic based
modeling and control design
• Workload parameterized dynamic
models and controller structure
• On-line optimization &
implementation
Infrastructure of the Networked Robotics
& Signal Intelligence Laboratory
Asok Ray, MNE Department
Overhead Vision based localization
Scanning Laser Rangefinder
Floor with embedded piezo-electric pressure sensors
Dedicated Wireless Network
NSF-ITR: An Agent-Based Negotiation Framework for the
Robust Design of Active-Passive Hybrid Piezoelectric
Vibration Control Networks – KW Wang, MNE Department
Distributed piezo-transducers and
multiple-branch circuitry networks with
active source and passive circuits
 Significantly increase smart
structure DOF and design space 
System much more adaptable to
advanced monitoring and control
Topology
Selection
Agent
Energy
Powerconcerned
Agent
FIRST LEVEL
FL-based
Negotiation
Agent
Robustness
Agent
Vibrationconcerned
Agent
Robustness requirement
SECOND LEVEL
Selected
Topology
GA-based
Sensitivity
Agent
Weights
GA-based
Optimization Agent
Design variables
Fitness
THIRD LEVEL
Models of
Different
Topology
Solve Ricatti /
Lyapunov Equations
and Active Gains
Calculate Cost
Function J
Controller
V
V
V
Distributed
Circuitry
To Distributed Actuators
Advanced design framework Agent-based information
technology for robust design,
tradeoff negotiation and circuit
topology synthesis
 Aims to incorporate/learn
knowledge and heuristics to assist
designers to address high-level
smart structure design issues in a
systematic way
Cyber-infrastructure Grants
T. Simpson, MNE Department
Two NSF CI-TEAM Awards
 Prof. Tim Simpson is leading two multi-university NSF CI-TEAM1 projects to
develop a cyber-infrastructure-based network of design repositories to support
innovative education and instruction in product design and development
 MNE, IME, SEDTAPP, IST, and Psychology faculty involved at Penn State
 Partner institutions include:
1. Northwestern University
2. Bucknell University
3. Drexel University
4. Virginia Tech
5. SUNY-Buffalo
6. Sweet Briar College
7. Norfolk State University
8. University of Missouri-Rolla
Repository
(Penn State)
Repository
(Drexel)
Repository
(VaTech)
Partnered
Repositories
1CI-TEAM
= Cyber-infrastructure Training, Education, Advancement, and Mentoring Program
SUMMARY

Penn State is continuing to build leading
educational and research programs.

A significant initiative in the College -network focused programs.