CSE Bits and Bytes Magazine

CSE Bits and Bytes

Could Woolly Mammoths Roam the Earth Again?

Fall 2010The Magazine of the Penn State Department of Computer Science and Engineering

A CSE professor believes that it is possible after helping to unlock the complete genetic makeup of the extinct animal.

From the Department Head

Greetings from Happy Valley! Since the last newsletter, we had an outstanding year.

Webb Miller, along with his colleague was selected as the Time Magazine’s 100 most influential people for their research on the woolly mammoth. Tom La Porta and colleagues were awarded the $35.5 million U.S. Army Center for Networking. President Barack Obama awarded Adam Smith and Sean Hallgren the prestigious National Science Foundation PECASE Award. Mary Jane Irwin was inducted into the Academy of Arts and Sciences.

Our faculty are continuing to do cutting edge research and publishing in exclusive outlets in such varied fields ranging from cryptography, quantum computing, privacy, genomics, bioinformatics, 3D VLSI, multi-core architecture, computer vision, cloud computing, sustainability, machine learning, next generation networking and security, and cyber science. This is reflected in Penn State Computer Science research expenditures to be ranked 6th nationally.

Our faculty and students are working with colleagues from industry, government, and academia. Our students are learning computing fundamentals mixed with latest technology and practice working along with our industrial partners. The Wall Street Journal recruiter rankings placed Penn State Computer Science 7th in the country.

Patrick Traynor, a Ph.D. alumnus of Penn State, who is now faculty at Georgia Tech, received the prestigious NSF Career Award. Patrick Roche, a 1975 computer science alumnus, was chosen as a 2010 Penn State Outstanding Engineering Alumnus.

On behalf of the CSE department, I would like to commend the excellent endeavors of our faculty, staff, students, and alumni. We are all excited to be part of a highly ranked and regarded program.

Phot

o by

Jen

Latc

hfor

d

CONTENTS

Faculty Research

Mean, Lean, and Green Data Centers

Smart-Meter Security Research

Department News

Mary Jane Irwin Elected as Fellow of the American Academy of Arts and Sciences CSE Selected by the Department of Energy for a $3 Million Project for Future Exascale Computer System Design

Five CSE Faculty Members Part of the Organizing Team for HPCA-16

Connect to CSE with Social Media

Features

Webb Miller is One of the First Scientists to Sequence Woolly-Mammoth Genome

Two CSE Faculty Members Receive Prestigious Presidential Early Career Award for Scientists and Engineers

Penn State Named Leader of New Network Science Collaborative Technology Alliance Communication Networks Research Center

Alumni

Patrick B. Roche Named Penn State Outstanding Engineering Alumnus

Family Members Honor James F. Kelly by Sponsoring the “James F. Kelly” Distinguished Lecture Series at CSE

Former CSE Student, Patrick Traynor, Receives NSF CAREER Award

Bits and Bytes Magazine is published once a year by the Penn State Department of Computer Science and Engineering for alumni, faculty, staff, students, and friends.

Department Head | Raj Acharya Editor and Designer | Jen Latchford ([email protected])

Photography | Cover illustration courtesy of ExhibitEase LLC-Steven W. Marcus

The Pennsylvania State UniversityDepartment of Computer Science and Engineering342 Info. Sciences and Technology BuildingUniversity Park, PA 16802Telephone 814-865-9505

Copyright, 2010 by the Penn State Department of Computer Science and Engineering. Penn State is committed to affirmative action, equal opportunity, and the diversity of its workforce. This publication is available in alternative media upon request.

U.Ed. ENG 11-22

1 2

3

4

5 7

10

13

15

17

18

20

1 | CSE Bits and Bytes | fall 2010

Mary Jane Irwin, Evan Pugh Professor and A. Robert Noll Chair in Engineering in the Department of Computer Science and Engineering was elected a fellow of the American Academy of Arts and Sciences for 2009. She is among the 210 new fellows and 19 foreign honorary members in the sciences, the humanities and the arts, business, public affairs and the nonprofit sector to be honored.

The academy, established in 1780 by the Founding Fathers, undertakes studies of complex and emerging problems. Current projects focus on science, technology and global security; social policy and American institutions; the humanities and culture; and education. The academy’s membership of scholars and practitioners from many disciplines and professions gives it a unique capacity to conduct a wide range of interdisciplinary, long-term policy research.

Mary Jane Irwin has been on the faculty at Penn State since 1977. Dr. Irwin received her Ph.D. from the University of Illinois and an Honorary Doctorate from Chalmers University, Sweden. She was named a Fellow of The Institute of Electrical and Electronic Engineers (IEEE) in 1995, a Fellow of The Association for Computing Machinery in 1996, and was inducted into the National Academy of Engineering in 2003. Awards she has received include the 2003 IEEE/CAS VLSI Transactions Best Paper of the Year Award, the 2004 DAC Marie R. Pistilli Women in EDA Award, the 2005 ACM Distinguished Service Award, the 2006 Computing Research Association Distinguished Service Award, the 2007 Anita Borg Technical Leadership Award, and the 2010 ACM Athena Lecturer Award.

“These remarkable men and women have made singular contributions to their fields, and to the world,” said Academy President Emilio Bizzi. “By electing them as members, the academy honors them and their work, and they, in turn, honor us.”

DEPARTMENT NEWSPh

oto

by C

urti

s Cha

n

Mary Jane Irwin Elected Fellow of the American Academy of Arts and Sciences

Penn Sate | www.cse.psu.edu | 2

Yuan Xie, associate professor of computer science and engineering, along with collaborators from Oak Ridge National Lab, Hewlett-Packard (HP), and the University of Michigan, have been selected by the Department of Energy (DOE) for a $3 million, three-year project titled “Advanced Architectures and Critical Technologies for Exascale Computing.”

Many scientific challenges (such as understanding the causes and potential impacts of climate change, enabling the mapping of the human genome, and a variety of national security challenges) require computational capabilities at extreme scale, moving from today’s Petascale computing to the future Exascale computing. In computing, petascale refers to a computer system capable of reaching performance in excess of one petaflop, i.e. one quadrillion floating point operations per second. Exascale computing is a thousand-fold increase over the petascale computing. It is likely that the future exascale computer systems will be comprised of as many as a billion microprocessor cores. According to the DOE, projections suggest that building an exascale computer from today’s technology would cost approximately $1 billion, with power consumption of more than a gigawatt and a mean time between failures of only ten minutes. Consequently, “disruptive technologies” that can help to address the challenges of building feasible future exascale computing systems are urgently needed.

Dept. of Energy Awards $3 Million Project for Future Exascale Computer System Design

continued on page 4


High Performance Computer Architecture (HPCA) is a well-known premier conference in the field of computer architecture. It provides a high-quality forum for scientists and engineers to present their latest research findings on all aspects of high-performance computer architecture. HPCA-2010 was the 16th edition of this series and the first major computer architecture conference to be hosted in India. HPCA-16 was held in Bangalore, the silicon city of India, Jan. 9-14, 2010. This year’s conference had 36 high-quality papers from industry and

academia including ten workhops and tutorials on emerging topics in computer architecture area.

The organizing team of HPCA comprised of many eminent personalities from both academia and industry. Organizing team members from Penn State consisted of Chita R. Das, Vijaykrishnan Narayanan, Yuan Xie, Anand Sivasubramaniam, and Raj Acharya.

—Asit Kumar Mishra

Five CSE Faculty Members Part of the Organizing Team for HPCA-16

This photo shows some of the HPCA-16 organizing members (including Anand Sivasubramaniam, Raj Acharya, and Chita R. Das) posing with the former president of India, A.P.J. Abdul Kalam, who inaugurated the conference.

DEPARTMENT NEWS


Xie, whose research has been recognized for the pioneer work in design techniques for emerging three-dimensional stacked integrated circuits (3D ICs) and emerging non-volatile memory (NVM)technologies, has teamed up together with HP’s Exascale Computing Lab, Oak Ridge National Laboratory, and University of Michigan, and proposed to identify and evaluate the most promising hardware and software solutions centered on the 3D IC and NVM technologies that can

Become a Fan of CSE on FacebookIt seems that everyone these days is jumping on the social media bandwagon, and CSE is no exception. We’ve recently launched a Facebook page:

www.facebook.com/psucse

We use Facebook to communicate our latest news and events to an audience of students, alumni, faculty, staff, and other friends.

Stay connected to CSE by joining our Facebook page. This networking community offers an opportunity to build and strengthen your professional contacts. You may also want to consider joining the Penn State College of Engineering Facebook group (www.facebook.com/psuengineering).

So, join us online! We look forward to hearing from you!

Dept. of Energy continued from page 2

address the reliability, energy, and performance requirements in the future exascale systems.

Xie has been a faculty member in the CSE department since 2003. He is a recipient of several awards including the NSF Career award and an IBM Faculty Fellowship.

—Yuan Xie and Jenny Latchford

Members of the Computer Systems Laboratory including Bhuvan Urgaonkar (left), Sriram Govindan, Byung Chul Tak, Anand Sivasubramaniam, Di Wang, Matt Derring, Aayush Gupta


The increasing reliance on information technology (IT) in large enterprises, and the growing popularity of Internet services including e-mail, e-commerce, and social networking, has proliferated computing and storage resources. The growing scale and complexity of these resources warrant a level of sophistication in their design, deployment, and management, mandating significant capital, operating and human resource expenditure. Consolidating these resources physically, either within an enterprise or outsourced to an external “cloud,” to leverage economies of scale has become critical to lowering this expenditure. This has led to a rapid growth in the number and size of data centers across enterprises of all scale and backgrounds.

Building a data center to house the servers, storage, and network switches is an expensive proposition. Compared to typical residential or commercial building costs that run to a few hundred dollars per square foot, construction costs for data centers with their raised floors, high-power densities, and extensive ducting to support wiring and cooling can run to several thousand dollars per square foot. One needs to amortize these costs by stretching the lifetimes of these data centers, ensuring they can accommodate the IT growth over the years while continuing to operate efficiently over their lifetime. Fortunately, hardware

innovations have allowed us to reduce form-factors at all levels - billions of transistors per chip, several blade servers and disks per chassis, etc. This allows us to progressively pack in more compute power and storage capacity to accommodate growing needs within the precious data center real-estate. However, the consequence of this miniaturization is a high power density that must be sustained to allow reliable operation.

Power consumption of data centers has come under much scrutiny recently, owing to cost, design complexity, and ecological concerns. The Penn State Computer Systems Laboratory (CSL), managed by Anand Sivasubramaniam and Bhuvan Urgaonkar, has been at the forefront of numerous innovations addressing a range of important problems related to datacenter energy and peak power optimization.

Powering a server over its four- to five-year lifetime can run to almost 50 to 100 percent of the initial hardware capital expenditure of the server. CSL has made notable contributions to reducing the energy consumption of data centers by developing mechanisms for modulating the hardware knobs – power modes and sleep states – exposed by today’s

FACULTY RESEARCH

Mean, Lean, and Green Data Centers

Phot

o by

Jen

Latc

hfor

d

Members of the Computer Systems Laboratory including Bhuvan Urgaonkar (front left), Anand Sivasubramaniam, Sriram Govindan, Byung Chul Tak, Di Wang (back left), Aayush Gupta, and Matt Derring.


servers in order to meet the service level agreements that data centers need to meet for a given workload. CSL has also been a pioneer in examining the implications of power expended by disk drives and proposing a novel solution of dynamic RPM modulation as a hardware knob for storage systems. Dynamic RPM modulation has since become a popular approach for building “green” disk drives.

Another dimension to the power problem is the capital and operating expenditures that power consumption imposes on IT equipment, and on its supporting infrastructure. Even if the overall energy consumption is not a concern, high peak power draw dissipates as heat and requires efficient cooling infrastructure to ensure reliable operation of the solid-state, electrical, and mechanical devices in the data center. A pre-requisite to optimizing the placement and optimization of the devices, and the cooling infrastructure, is a thermal modeling tool for answering “what-if ” questions. CSL has developed a rack-level thermal modeling tool called ThermoStat, that is being distributed in the public domain.

Even if it is a rare event, peak power draw impacts the monthly electricity bill as well. There are two components to the commercial tariff imposed by a utility – one based on the units (Kwh) consumed over the month, and another is the cost per maxi-mum kilowatt, typically calculated as an average over a duration such as 15 minutes, at any time during the month. The second component can contribute to a large portion of the bill, especially when theaverage power draw is much lower than the occasional peak. This is because peak draw directly impacts generation capacity, and we are already close to the limits of current generation capacities as noted by the Environmental Protection Agency. The peak draw also determines the provisioning capacities for the grid, even if this draw is not sustained. Time-of-day and other dynamic pricing mechanisms are employed by utility providers to influence the peak demands of their consumers. At the consumer end – typically throttling by either shutting down servers or employing lower power consuming states in the datacenter – is the only option for demand shaping of the peak power draw. CSL has come up with an entirely novel approach to this problem by proposing the use of uninterruptible power supply (UPS) batteries, already present in data centers for peak shaving. UPS batteries are primarily used as a back-up power source upon electricity outages, and are very rarely called upon in the realworld, and even in those rare cases they are temporary (a few seconds to at most a few minutes) fail-over vehicles

until the diesel generators are activated. By using valleys in power consumption to charge the batteries which can then be used to offset some of the draw from the utility during peaks in the workload, allows computing and storage devices to operate as is (unlike throttling) while still optimizing the power bill. CSL researchers have shown that this is a viable option for today’s batteries used in UPS devices, without compromising on the lifetime of these devices or the availability of the datacenter. These energy storage devices/buffers will play an increasing role going forward, as the cost, availability and carbon footprint of the energy source become more dynamic.

In addition to research papers, Sivasubramaniam has been involved in reducing the power consumption of two production data centers of a large IT services company by around 20 percent. More than 15 Ph.D. students have graduated from CSL in the past decade, with two going to tenure-track positions in top U.S. universities and several others to research labs at leading companies. The research has been generously funded over the years by the National Science Foundation, and grants from IBM, HP, Microsoft, Cisco, and Seagate.

—Anand Sivasubramaniam


Society is becoming increasingly aware of the security and privacy risks associated with everyday technologies such as e-commerce and social networking. A recent shift towards smart electric grids will introduce one new such risk that many people will be unfamiliar with, smart electric meters [1].Smart-meters are replacing current electromechanical meters to provide services like automatic meter reading, fine-grained usage data, and the ability for utilities to remotely disconnect power from defaulters. To do this, smart-meters will introduce Internet-connected embedded systems into the home of every electricity consumer in the world as shown in figure 1. Given the power delegated to these meters and the speed with which they are being deployed, it is important that we develop and understanding of the potential risks they pose now rather than later.

The Systems and Internet Infrastructure Security (SIIS) lab, co-directed by Patrick McDaniel and

Trent Jaeger, both associate professors of computer science and engineering, is studying smart-meter security problems and mitigations. This study spans what have emerged as the three principal security concerns for smart-metering. (1) Electricity generators and distributors are worried about losing revenue due to energy theft by hacked meters. (2) Customers are concerned with their privacy in the face of meters that can determine their appliance usage patterns based on the power drawn over time. (3) Governments are worried about the potential for large-scale abuse of meter remote-disconnect features to cause widespread blackouts.

In a joint project sponsored by Lockheed Martin, the SIIS Lab is performing black box penetration testing of smart-meters to find security issues that are of concern to electric utilities, the paramount of which is theft of service. The lab’s findings show that energy theft is not only feasible, but highly likely

Smart-Meter Security Research

Dmitry Podkuiko (left) and Sergei Miadzvezhanka display the smart-meter pen testers.

FACULTY RESEARCH

continued on page 9


in the advanced metering infrastructure. Fraud with electromechanical meters requires dangerous physical tampering with the small payoff of a single compromised meter. Smart-meter tampering, however, can reveal vulnerabilities or passwords that can be redistributed via the Internet to customers willing to take the small risk of a network-based attack against their own meter.

The attack tree in figure 2 shows a number of potential scenarios for committing energy theft in AMI. In one such scenario, the SIIS team found

a vulnerability in the collector meter that acts as a gateway between meters and the telephone network that allows a malicious user to modify billing data on the wire. In another, similar scenario, it was found that by extracting a deployment-wide meter password directly from a meter’s EEPROM, it is possible to commit fraud from any meter in the deployment. The potential profit from the distributions of such leaked passwords represents an economic incentive for meter tampering.

The lab is continuing to work with Lockheed Martin and the meter vendors to disclose fraud issues, which currently have been publicized without vendor information in [2,5].

While utilities are concerned with protecting their bottom line, privacy advocates are aware of the potential abuse of fine-grained meter data. Previous generation meters could only display monthly net totals of energy usage. Smart-meters however, can record a minute-by-minute load profile of the instantaneous power usage in the house. The load profile contains edge events

Figure 1. Smart-meter network topologies vary by deployment and vendor.

Figure 2. Attack trees are an example of the tools used to pinpoint vulnerabilities in smart-meters.

corresponding to on and off events of large appliances. This appliance-level information has been shown to reveal characteristics of occupant behavior [3].

A well-studied class of algorithms called Non-Intrusive Load Monitors use clustering methods to match events in load profiles with particular appliances such as those related to cooking, air conditioning, entertainment, etc. To reduce the amount of information that meters can extract from load profiles, the SIIS team is developing a method of Non-Intrusive Load Levelling (NILL) that uses

energy storage to level large load events. As an example, a NILL algorithm can schedule an in-home solar battery to charge throughout the day while the occupants are at work, and discharge to supply part of the load exerted by appliance use when the occupants return home. A smart-meter attempting to find appliances used to cook dinner will be unsuccessful, as their characteristic loads will be partially supplied by the battery.

The final area of smart-meter security being studied by the SIIS lab is what many consider

to be the worst- case scenario for a cyber attack against the electric grid. An adversary possessing a remote meter exploit could program thousands to millions of meters to cut off power to customers at a strategic time, before rendering themselves inoperable. This is a plausible scenario, because smart-meters deployments are mono-cultures, collections of


network nodes that run identical software stacks. An exploit discovered by inspection of a single physical meter is applicable to all meters in the monoculture, making it highly likely that eventually, someone will discover a vulnerability allowing for a blackout attack.

A known mitigation strategy for widespread exploitation of monocultures is the use of software diversity. Software diversity techniques add a degree of uniqueness to a program’s implementation without modifying its interface. For example, a diversity technique known as Address Space Layout Randomization (ASLR) randomizes the base address of a program’s code segment to foil exploit attempts that leverage existing program code for malicious purposes. Unfortunately, most embedded processors used in smart-meters do not support the memory management units needed for techniques like ASLR. Other common mitigation strategies including stack cookies and non-executable stacks are also not widely supported.

Students in the SIIS lab developed a redundant address encryption scheme to protect function return addresses from exploitation [4].

The scheme uses multiple encryptions of a return address to both prevent return address exploitation, and to prevent random program errors due to an undetected failed exploit.

The lab’s future efforts towards security smart electric grids will include fraud detection in demand response systems where customers can obtain rebates by curtailing usage during peak load, and analysis of the upstream impact of meter exploits on core grid components.

[1] P. McDaniel, S. McLaughlin. 2009. Security and Privacy Challenges in the Smart Grid. IEEE Security & Privacy Magazine 7(3):75-77.

[2] S. McLaughlin, D. Podkuiko, P. McDaniel. Sept. 2009. Energy Theft in the Advanced Metering Infrastructure. Proceedings of CRITIS 2009. Bonn, Germany.

[3] M. A. Lisovich, D. K. Mulligan, S. B. Wicker. 2010. Inferring Personal Information from Demand-response Systems. IEEE Security & Privacy 8(1):11-20.

[4] S. McLaughlin, D. Podkuiko, A. Delozier, S. Miadzvezhanka, P. McDaniel. Aug. 2010. Embedded Firmware Diversity for Smart Electric Meters. Proceedings of HotSec 2010. Washington, D.C.

[5] S. McLaughlin, D. Podkuiko, S. Miadzvezhanka, A. Delozier, P. McDaniel. Dec. 2010. Multi-vendor Penetration Testing in the Advanced Metering Infrastructure. To appear in Proceedings of ACSAC 2010. Austin, TX.

—Patrick McDaniel andStephen McLaughlin

FACULTY RESEARCH

Smart-Meter Security continued from page 8


Scientists at Penn State are leaders of a team that is the first to report the genome-wide sequence of an extinct animal, according to Webb Miller, professor of biology and computer science and engineering and one of the project’s two leaders. The scientists sequenced the genome of the woolly mammoth, an extinct species of elephant that was adapted to living in the cold environment of the northern hemisphere. They sequenced four billion DNA bases using next-generation DNA-sequencing instruments and a novel approach that reads ancient DNA highly efficiently. More information about this project is on the Web at http://rw.mammoth.psu.edu/.

“Previous studies on extinct organisms have generated only small amounts of data,” said Stephan C. Schuster, Penn State professor of biochemistry and molecular biology and the project’s other leader. “Our dataset is

100 times more extensive than any other published dataset for an extinct species, demonstrating that ancient DNA studies can be brought up to the same level as modern genome projects.”

The researchers suspect that the full woolly mammoth genome is the size of the modern-day African elephant’s genome. Although their dataset consists of more than 4 billion DNA bases, only 3.3 billion of them – a little over the size of the human genome – currently can be assigned to the mammoth genome. Some of the remaining DNA bases may belong to the mammoth, but others could belong to other organisms, like bacteria and fungi, from contaminated the sample. The team used a draft version of the

continued on page 11

Phot

o by

M. S

cott

John

son

Webb Miller (left) and Stephan Schuster

Scientists Sequence Woolly Mammoth Genome

FEATURES


African elephant’s genome, which currently is being generated by scientists at the Broad Institute of MIT and Harvard, to distinguish those sequences that truly belong to the mammoth from possible contaminants.

“Only after the genome of the African elephant has been completed will we be able to make a final assessment about how much of the full woolly mammoth genome we have sequenced,” said Miller. The team woolly mammoth’s genome when the project receives additional funding.

The team sequenced the mammoth’s nuclear genome using DNA extracted from the hairs of a mammoth mummy that had been buried in the Siberian permafrost for 20,000 years and a second mammoth mummy that is at least 60,000 years old. By using hair, the scientists avoided problems that have bedeviled the sequencing of ancient DNA from bones because DNA from bacteria and fungi, which always are associated with ancient DNA, can more easily be removed from hair than from bones. Another advantage of using hair is that less damage occurs to ancient DNA in hair because the hair shaft encases the remnant DNA like a biological plastic, thus protecting it from degradation and exposure to the elements.

The researchers previously had sequenced the woolly mammoth’s entire mitochondrial genome, which codes for only 13 of the mammoth’s roughly 20,000 genes but is relatively easy to sequence because each of the mammoth’s cells has many copies. In their most recent project, the team sequenced the mammoth’s nuclear genome, which codes for all the genetic factors that are responsible for the appearance of an organism. The two methods combined have yielded information about the evolution of the three known elephant species: the modern-day African and Indian elephants and the woolly mammoth. The team found that woolly mammoths separated into two groups around two million years ago, and that these groups eventually became genetically distinct sub-populations. They also found that one of these sub-populations went extinct approximately 45,000 years ago, while another lived until after the last ice age,

about 10,000 years ago. In addition, the team showed that woolly mammoths are more closely related to modern-day elephants than previously was believed.“Our data suggest that mammoths and modern-day elephants separated around 6 million years ago, about the same time that humans and chimpanzees separated,” said Miller. “However, unlike humans and chimpanzees, which relatively rapidly evolved into two distinct species, mammoths and elephants evolved at a more gradual pace,” added Schuster, who believes that the data will help to shed light on the rate at which mammalian genomes, in general, can evolve.

The team’s new data also provide additional evidence that woolly mammoths had low genetic diversity. “We discovered that individual woolly mammoths were so genetically similar to one another that they may have been especially susceptible to being wiped out by a disease, by a change in the climate, or by humans,” said Schuster. While members of the team previously ruled out humans as a cause of extinction for at least one of the Siberian sub-populations the group appears to have gone extinct at least 45,000 years ago at a time when there were no humans living in Siberia – much debate still remains regarding the causes of extinction for the other group and for those populations that lived in other places, such as North America.

Currently, the team is searching the mammoth’s genome for clues about its extinction. “For example,” said Miller, “most animal genomes contain integrated viral sequences and, though these are not directly associated with disease, evidence of multiple recent integration events could indicate a perturbation of virus-host interaction that might be responsible for disease. Alternatively, it might turn out that long generation times and limited outbreeding result in accumulation of deleterious genetic mutations. We are considering a number of possible causes of extinction.”

The new data are allowing the Penn State team to begin looking for genetic causes of some of the mammoth’s unique characteristics, such as their adaptation to extremely cold environments. For instance, the team already has identified a number of

FEATURES

Woolly Mammoth Genome continued from page 10


cases in which all previously sequenced mammals, except mammoths, have the same protein segment. “One has to wonder whether a particular protein that has remained the same in animals for several billion years of combined evolution and then became different in mammoths could result in a mammoth-specific trait,” said Miller.

Investigating the unique characteristics of woolly mammoths and why they went extinct are just some of the many tasks that the research team plans to pursue now that they have access to such a large quantity of sequence data. “This really is the first time that we have been able to study an extinct animal in the same detail as the ones living in our own time,” said Schuster.

Another significant aspect of the study is that it was completed by a small group of scientists at a relatively low cost and over a short period of time, whereas previous reports of modern mammalian genome sequences – including human sequences – have taken millions of dollars and several years of analysis by large groups of scientists to complete. Miller hopes that after he completes a few additional genome projects he can produce computer software that will enable others to perform low-cost mammalian genome analysis, and Schuster already is preparing to decode extinct genomes at an even faster pace.Schuster hopes that lessons learned from the mammoth genome about why some animals go extinct while others do not will be useful in protecting other species from extinction, such as the Tasmanian

devil, whose survival is threatened by a deadly facial cancer. “In addition,” added Schuster, “by deciphering this genome we could, in theory, generate data that one day may help other researchers to bring the woolly mammoth back to life by inserting the uniquely mammoth DNA sequences into the genome of the modern-day elephant. This would allow scientists to retrieve the genetic information that was believed to have been lost when the mammoth died out, as well as to bring back an extinct species that modern humans have missed meeting by only a few thousand years.”

In addition to being members of the faculty of Penn State’s Department of Biochemistry and Molecular Biology, Miller and Schuster are researchers associated with Penn State’s Center for Comparative Genomics and Bioinformatics. The study also involved researchers from the Severtsov Institute of Ecology and Evolution and the Zoological Institute in Russia, the University of California, the Broad Institute, the Roche Diagnostics Corp. and the Sperling Foundation in the

United States. This research was funded by Penn State, Roche Applied Sciences, a private sponsor, the National Human Genome Research Institute and the Pennsylvania Department of Health.

—Barbara Kennedy

Photos by Stephan Schuster and Colleagues


Sean Hallgren and Adam Smith, associate professors of computer science and engineering, have been named as recipients of the highest honor that any young scientist or engineer can receive in the United States, a Presidential Early Career Award for Scientists and Engineers (PECASE). Hallgren and Smith were among 100 researchers nationwide, and two of 20 National Science Foundation (NSF) nominees, to receive this national award.

Hallgren works in the area of quantum computation, which aims to use quantum mechanical systems for computation. Quantum computers can break widely used cryptosystems, including those used to

protect e-commerce transactions. His work aims to find new applications for quantum computers and to determine which cryptosystems are secure against them.

Smith studies cryptography and information privacy and their connection to such diverse fields as quantum mechanics, combinatorics, information theory, and statistics. His CAREER award focuses on the problems stemming from conflicts between data access and privacy in collections of personal and sensitive data such as census surveys, social networks, and public health data. His work addresses the need for formal privacy guarantees that remain

Two CSE Faculty Members Receive Prestigious PECASE Award

FEATURES


science and engineering, states that “receiving an NSF PECASE is a very exclusive recognition. The Computer Information Science and Engineering (CISE) division typically nominates three of the available twenty slots to NSF every year. The CSE department at Penn State received two of the three CISE awards. It is very unusal to have two PECASE awards recipients in the same department.”

Hallgren and Smith were nominated for the PECASE by NSF.

meaningful even against an intruder with partial knowledge of the sensitive data. Smith says that “he is very honored by the award. He thinks that a share of the recognition that comes with the PECASE should go to his mentors and scientific colleagues with whom he has been fortunate to collaborate.”

The PECASE program recognizes outstanding scientists and engineers who, early in their careers, show exceptional potential for leadership at the frontiers of knowledge. This Presidential Award is the highest honor bestowed by the U.S. government on scientists and engineers beginning their careers.Raj Acharya, professor and head of computer

Two CSE Faculty Members Receive Prestigious PECASE Award

Offi

cial

Whi

te H

ouse

Pho

to b

y La

wre

nce

Jack

son

Phot

o by

Dav

id K

amm


In the photo on the previous page from left: Will Leland, Network Science Collaborative Technology Alliance program director; Tom La Porta, Tad Elmer, president, Raytheon BBN Tech.; Maj. Gen. Nickolas Justice, commanding officer of U.S. Army Research, Development and Engineering

Command; Alexander Kott, Applied Research Laboratory Network Science Collaborative Technology Alliance Collaborative agreement manager; Raju Namburu, Gary Blohm, director, U.S. Army Communications-Electronics Research, Development and Engineering Center.

Penn State Named Leader of New Network Science Collaborative Technology Alliance Communication Networks Research Center

Penn State will launch a new interdisciplinary research center focused on the science of communication networks with approximately $35.5 million in funding over ten years from the Army Research Laboratory.

The Communication Networks Research Center, part of the new Network Science Collaborative Technology Alliance, will perform foundational research on network science. The research focuses on the interplay among the social/cognitive,

information and communication networks.Thomas La Porta, distinguished professor of computer science and engineering and director of the Networking and Security Research Center (NSRC) at Penn State, is the principal investigator and director of the Communication Networks Research Center. Other NSRC faculty participating in the new center include Guohong Cao, professor of computer science and engineering; Adam Smith, assistant professor of computer science and engineering; Sencun Zhu, assistant professor of

FEATURES


computer science and engineering; and Aylin Yener, associate professor of electrical engineering.

Partners in the center include the University of California at Davis, the University of California at Santa Cruz, University of Southern California and the City University of New York. Several institutions also will be collaborating with the center including the University of California at Riverside, North Carolina State University, Stanford University and BBN Technologies.

“This is the first project looking at the social interaction, information distribution and mobile ad-hoc network as a whole,” explained Jay Gowens, director of the Army Research Laboratory’s Computational and Information Sciences Directorate (CISD). “This comprehensive approach will develop network science as an integrated whole, combining three previously disparate intellectual pursuits.”

Alexander Kott, network science division chief at the CISD, said, “The Army is moving rapidly and ever deeper into a network-centric world,” said Alexander Kott, network science division chief at the CISD. “So much now depends on how war fighters and sensors and weapons communicate information through mobile, self-forming rapidly-changing networks.”

Eva Pell, former senior vice president for research and dean of the Graduate School, said, “We have been building a strong research effort in network sciences at Penn State. The Networking and Security Research Center at Penn State is uniquely position to lead this work.”

The goal of the Communication Networks Research Center is to understand the why a communication network, like the Internet or a wireless network, behaves over time.

“We will focus on unique ways in which communication networks interact with social networks and information networks,” said La Porta. “For example, we can all see how new ways of building social networks through services such

as Facebook impact how communication networks like the Internet and cellular phone networks are used. Ultimately we will be able control the behavior of communication networks in a way that allows people to exchange the most important information.”

—Curtis Chan

Phot

o by

Pau

l Haz

i


Patrick Roche Named Penn State Outstanding Engineering AlumnusPatrick Roche is one of the recipients of the 2010 Penn state Outstanding Engineering Alumni Award. The award is the highest honor conferred by the College of Engineering and recognizes graduates who have received exceptional levels of professional achievements.

During his sophomore year at Penn State, math major Pat started thinking about what he was going to do when he got out of school. He ended up taking a computer programming course and was immediately hooked. “I thought it was really fun, and if I could get paid to do it, that would be a bonus!”

Pat went on to double-major in math and computer science, earning bachelor’s degrees in both fields. When he graduated, however, the job market was down, so he decided to stay at Penn State and earn his masters. “My goal was to work at Bell Laboratories,” he explains, referring to the research and development organization that was then part of AT&T. “At that point they weren’t hiring, so I thought maybe I’d have a better chance if I had my master’s degree.”

Pat’s strategy worked out well. After receiving his master’s degree in computer science, he joined the technical staff at Bell Labs. He spent the first three years of his career as a developer on various operations support systems with responsibility for architecture, design, and coding. By 1980, he was leading software development groups involved in the rollout of new products, including those that performed work coordination, automatic circuit testing, and workforce administration. Working at Bell Labs gave Pat the opportunity to brainstorm about his projects with some of the world’s most eminent scientists. “There were experts in physics,

ALUMNI


Family Members Honor James F. Kelly by Sponsoring a Distinguished Lecture Series

The Department of Computer Science and Engineering held its first ever sponsored distin-guished lecture series. Family members sponsored the “James F. Kelly” Distinguished Lecture Series. Barbara Liskov was selected as the speaker.

Liskov is an Institute Professor at MIT and also associate provost for faculty equity. She is a member of the National Academy of Engineering, a fellow of the American Academy of Arts and Sciences, and a fellow of the ACM. She received the ACM

Turing Award in 2009, the ACM SIGPLAN Programming Language Achievement Award in 2008, the IEEE Von Neumann medal in 2004, a lifetime achievement award from the Society of Women Engineers in 1996, and in 2003 was named one of the 50 most important women in science by Discover Magazine. Her research interests include distributed systems, replication algorithms to provide fault-tolerance, programming methodology, and programming languages.

continued on page 19

mathematics, and computer science—especially computer science,” he recalls. “It was amazing to me that I could go talk to the guys that invented the Unix operating system, for example. It was just a great place.”

In 1984, Bellcore was spun off from Bell Labs; it was renamed Telcordia Technologies in 2002. Pat went on tospend 31 years there, advancing to various leadership posi-tions related to software development, including executive director, general manager, group senior vice president, and senior vice president. He also had the opportunity to work internationally,

collaborating with partners and customers across the globe. One of his biggest challenges, he says, was learning how to pick teams that would work well together and how to encourage the teaming required for success. “I found out that if you always take the best people, sometimes you don’t get anything done because they spend all their time trying to prove they’re smarter than everyone else,” he laughs.

Pat retired in 2008. He and his wife, Cynthia ('74 HHD), live in Hillsborough, NJ.

—Jane Harris


Jim Kelly graduated with a bachelor of science in electrical engineering from the Penn State College, on Jan. 27, 1953. Jim always said that Penn State gave him the chance he needed in life to start a career and a family. As the son of a coal miner in Dunmore, PA, Jim was the first one of his family to go to college. He always remarked on the emphasis his parents placed on getting a college education and instilled this mindset into his children. Three of us are graduates of Penn State.

Jim met Marion in Scranton, PA, and the two were married in 1957. Dad integrated mom into Penn State in 1959 taking her to their first football game. Traffic was a problem on game day, as they arrived late. Unfortunately, the Lions lost that day, but that didn’t deter them from attending games and class reunions for the next 30 years. Beginning in 1982, Jim started a professional relationship with the University, launching projects with Electrical Engineering, Materials, and the Applied Research Laboratory. The rich tradition Jim set in motion continues today, as his granddaughter is starting at University Park this semester.

Jim was a life-long learner and dedicated to multi-disciplinary work. His motto was: “When someone tells me something can’t be done, I really get moti-vated and do it.” From an early beginning, he led a number of interesting projects as an engineer working for the United States government’s U.S. Army Material Command. One of his first projects was a major break-through in computer-controlled automatic test systems for performance and testing of circuits, assemblies, and complete systems. The project also provided for on-line generation, editing, and validation of test programs. The year was 1971. The EQUATE (AN/US-410) system started as a $1 million contract with RCA. By 1986, more than 150 systems were deployed with hundreds of thousands of written software test program sets. The wide spread adoption of EQUATE and its use across the military’s entire communications hardware suite was a testament to the impact of Jim’s ideas and his drive to use state-of-the-art in technology to solve operational problems.

Throughout his career, Jim focused on the latest in technology. Projects along the way included automation in epitaxial growth systems, automated control systems for the growth of silicon and GaAs boules, as well as, microwave device testing. Jim spent a great deal of time learning about new technology and how engineering analysis, measurement systems, computers and people could be linked to improve and miniaturize devices. His work on night-vision, data storage systems, laser welding systems, and coating processes all yielded advances in performance and manufacturing techniques.

Jim’s use of technology also extended to home. He was an early adopter of the personal computer, but rather than go with IBM, he chose the Mac. He bought his first Mac early, went to the local Penn State branch campus, enrolled in programming classes, purchased a compiler and went to work programming. Classic Jim, instead of buying the spreadsheet software for checkbooks, he wrote his own programs, in Fortran. Eventually he converted over, but he always felt you needed to know how to do it yourself. We all remember him sitting at the kitchen table working on math problems while we toiled over our own homework. As he said, you need to keep sharp, and besides, it’s fun. Jim was a wonderful engineer, a real Penn Stater, and by starting this speaker series, we honor his memory and the joy he found in his field.

—Chris Kelly

ALUMNI

James F. Kelly continued from page 10


Former CSE student, Patrick Traynor, received an NSF Faculty Early Career Development (CAREER) award. According to NSF, “This is the most prestigious award in support of junior faculty who exemplify the role of teacher-scholars through outstanding research, excellent education and the integration of education and research within the context of the mission of their organizations.” His proposal is titled, “CAREER: Protecting User Data on Lost, Stolen and Damaged Mobile Phones.”

Traynor is an assistant professor in the College of Computing at Georgia Tech. He earned his Ph.D. and M.S. in computer science and engineering at Penn State. He is currently a member of the Georgia Tech Information Security Center and a co-director of the Converging Infrastructure Security Laboratory.

His research focuses primarily on security in cellular networks. In particular, he explores the problems that arise as this piece of critical infrastructure is beginning to converge with the larger Internet. Traynor is also interested in the systems challenges of applied cryptography and security for the Internet, mobile devices and wireless systems. In general, he is curious in learning about how secure systems are designed, constructed and broken.

He was co-advised by Tom La Porta and Patrick McDaniel.

—Jen Latchford and Patrick Traynor

Former Student, Patrick Traynor, Receives NSF CAREER Award

Geo

rgia

Tec

h Ph

oto

by G

ary

Mee

k

Department of Computer Science and EngineeringThe Pennsylvania State University342 Information Sciences & Technology BuildingUniversity Park, PA 16802-6822

Nonprofit Org.U.S. PostagePAIDState College, PAPermit No. 1

CSE Bits and Bytes Magazine

Documents

Transcript of CSE Bits and Bytes Magazine