January

• 1/1/2007-1/2/2007 - - SDSC Officially Closed
  SDSC officially closed until January 3, 2007
  
  
• 1/11/2007 - TechTIPS - TechTIPS seminar
  Technology Transfer and Intellectual Property Services discussion, 5th floor Seminar Space, Atkinson Hall, UCSD at 12pm
  
  
• 1/12/2007 - Talk - Tom Anderson
  "Do Incentives Build Robustness in Distributed Systems?" at EBU3b (CSE) 1202 1-2PM
  
  

  Distinguished Lecturer Series Presents:

  "Do Incentives Build Robustness in Distributed Systems?"
  by Tom Anderson, University of Washington
  EBU3b 1202 1-2PM

  An emerging paradigm in the design of large scale distributed systems is to explicitly consider incentives as part of the design. In this talk, I outline two case studies of the role of incentives in distributed systems, in BGP and BitTorrent. In the case of BGP, we were able to show that detour routing inefficiencies in the Internet can be largely solved by better aligning incentives among ISPs. In the case of BitTorrent, a system explicitly designed to take incentives into account, we were able to show that its incentive mechanism can be easily subverted, that all users have an incentive to subvert the mechanism, and that the likely end result will be worse performance for all. In both cases I will outline solutions which are efficient, robust to incentives, and easy to implement.

  Tom Anderson?s research concerns the practical issues in constructing robust, secure, and efficient computer systems. He is a professor of Computer Science and Engineering at the University of Washington, where he also received his Ph.D. (1991) and M.S. (1989). His current research project is RIP (Re-architecting the Internet Protocols), a method for fixing the myriad problems with the Internet by re-thinking its design from first principles. http://www.cs.washington.edu/homes/tom

[Back to Calendar View]

February

• 2/5/2007 - Talk - Frank Kelly
  "Flow level models of Internet congestion control" hosted by Ruth Williams on Monday, February 5, 2007 at 12:00 P.M. in AP&M 6402.
  
  

  Frank Kelly from Cambridge is one of the best-known network systems researchers worldwide. He is a Fellow of the Royal Society and winner of the Koji Kobayashi and numerious other awards. He will speak Monday 2/5 in the Math department, hosted by Ruth Williams.

  Abstract: Variability in the number of simultaneous flows present can have a substantial impact on the perceived performance of packet networks such as the Internet. While the packet level behaviour of a given set of flows is by now well understood, less is known about the stochastic behaviour of the number of flows in progress on different routes through the network. In this talk we describe recent work on Brownian models of networks in heavy traffic. Joint work with Ruth Williams.

• 2/9/2007 - Talk - Athina Markopoulou
  "Optimal Allocation of Filters against DDoS Attacks" by Athina Markopoulou (UC Irvine) on Friday, February 9, 2007 at 11:00 A.M. in SDSC 408.
  
  

  Distributed Denial-of-Service (DDoS) attacks are a major problem in the Internet today. During a DDoS attack, a large number of compromised hosts send unwanted traffic to the victim, thus exhausting the resources of the victim and preventing it from serving its legitimate clients. One of the mechanisms that can be used against DDoS is filtering, which allows routers to selectively block unwanted traffic. Given the magnitude of DDoS attacks and the high cost of filters in the routers today, the successful mitigation of a DDoS attack using filtering crucially depends on the efficient allocation of filtering resources.

  In this work, we consider a single router, typically the gateway of the victim, with a limited number of available filters. We study how to optimally allocate filters to attack sources, or entire domains of attack sources, so as to maximize the amount of good traffic preserved, under a constraint on the number of filters. First, we look at the single-tier problem, where the collateral damage on legitimate traffic is high due to the filtering at the granularity of attack domains. Second, we look at the two-tier problem, where we have an additional constraint on the number of filters and filtering is performed at the granularity of attackers and/or domains. In both cases (i) we formulate and solve the problem of optimal filter allocation (ii) we propose efficient algorithms and (iii) we evaluate our solutions over a range of realistic attack-scenarios, based on data from the analysis of the Code Red and Slammer worms. Our results demonstrate that efficient filter allocation significantly improves the tradeoff between the number of filters used and the amount of legitimate traffic preserved.

  This is joint work with Karim El Defrawy from UC Irvine and Katerina Argyraki from EPFL, Lausanne.

  
  

  About the speaker:
  Athina Markopoulou is an assistant professor with the EECS Dept, UC Irvine. She received the Diploma degree in Electrical and Computer Engineering from the National Technical University of Athens, Greece, in 1996. She received the M.S. and Ph.D. degrees both in Electrical Engineering from Stanford, in 1998 and 2002 respectively. Prior to joining UCI, she worked at Stanford University, Sprint Advanced Technologies Labs, and Arastra, a startup, as postdoctoral research fellow, member of technical staff, and research scientist, respectively. Her research interests are in the area of networking, including Internet Denial-of-Service, network measurement and control, voice and video over IP networks.

• 2/15/2007 - Talk - Anupam Datta
  "Reasoning about Security and Privacy" talk given by Anupam Datta (Stanford University) on Thursday, February 15, 2007 at 12:00 P.M. in EBU3B 4140.
  
  

  ABSTRACT:
  
  My research focus is on developing principled methods for modeling, analysis, and design of security mechanisms that are applicable to real-world systems. In this talk, I will present two examples. In the first part, I will describe PCL - a logic for proving security properties of network protocols. This logic has been successfully applied to a number of internet, wireless and mobile network security protocols developed by the IEEE and IETF Working Groups, in several cases identifying serious security vulnerabilities. Two central results for PCL are a composition theorem and a computational soundness theorem. In contrast to traditional folk wisdom in computer security, the composition theorem allows proofs of complex protocols to be built up from proofs of their constituent sub-protocols. The computational soundness theorem guarantees that, for a class of security properties and protocols, axiomatic proofs in PCL carry the same meaning as reduction-style cryptographic proofshand-proofs. In the remaining time, I will describe my current work on LPU - a logic for stating and enforcing privacy policies and its application to problems in medical privacy. One significant contribution of this work is the reduction of problems pertaining to policy compliance, combination and refinement to standard decision procedures in temporal logic.

  BIOGRAPHY:
  
  Anupam Datta is a Research Associate in the Computer Science Department at Stanford University. He obtained PhD (2005) and MS (2002) degrees from Stanford University and a BTech from IIT Kharagpur (2000), all in Computer Science.

[Back to Calendar View]

March

• 3/7/2007 - Talk - David Kempe
  "On the Bias of Traceroute Sampling" (or: "Power-law Degree Distributions in Regular Graphs") talk given by David Kempe on Wednesday, Marth 7, 2007 at 3:00 P.M. in EBU3B 4109
  
  

  Abstract:

  Understanding the structure of the Internet graph is a crucial step for building accurate network models and designing efficient algorithms for Internet applications. Yet, obtaining its graph structure is a surprisingly difficult task, as edges cannot be explicitly queried. Instead, empirical studies rely on traceroutes to build what are essentially single-source, all-destinations, shortest-path trees. These trees only sample a fraction of the network's edges, and a recent paper by Lakhina et al. found empirically that the resuting sample is intrinsically biased. For instance, the observed degree distribution under traceroute sampling exhibits a power law even when the underlying degree distribution is Poisson.

  In this talk, we study the bias of traceroute sampling systematically, and, for a very general class of underlying degree distributions, calculate the likely observed distributions explicitly. To do this, we use a continuous-time realization of the process of exposing the BFS tree of a random graph with a given degree distribution, calculate the expected degree distribution of the tree, and show that it is sharply concentrated. As example applications of our machinery, we show how traceroute sampling finds power-law degree distributions in both d-regular and Poisson-distributed random graphs. Thus, our work puts the observations of Lakhina et al. on a rigorous footing, and extends them to nearly arbitrary degree distributions.

  (Joint work with Dimitris Achlioptas, Aaron Clauset, and Cristopher Moore.)

• 3/7/2007 - Talk - Dr. Jay Boisseau
  "Deploying the First NSF "Track 2" System: Technology Challenges and Opportunities" given by Dr. Jay Boisseau, Director of the Texas Advanced Computing Center, The University of Texas at Austin on Wednesday, March 7, 2007 from 4:15-5:00 P.M. at the Social Sciences Building, Room 108.
  
  

  Abstract: The Texas Advanced Computing Center (TACC) at The University of Texas at Austin was selected by the National Science Foundation (NSF) to deploy the first of at least four "Track 2" HPC systems. These systems will provide the path to petascale academic computing in the US, and will be followed by a "Track 1" system to be deployed in 2011.

  TACC's new system, designed by Sun, will have unprecedented CPU power, total memory, and disk and thus presents tremendous opportunities for computational researchers. However, the scale (more than 60,000 processor cores) and complexity of the system also present significant challenges for both users and administrators that must be overcome relatively quickly to maximize the impact of this system and NSF's investment in it.

  Challenges include: scalability of algorithms, tools and libraries; achieving high parallel applications performance on quad-core processors and 16-way nodes; managing a 3000+ node system reliably and securely; providing fault tolerance for applications and systems software; managing and analyzing data of unprecedented scale; exploring new programming methodologies and tools; and supporting development of multi-scale, multi-physics applications.

  Overcoming these challenges will require rapid and continuous advances, leveraging the best expertise in the nation (and worldwide). Thus, this system presents interesting opportunities for collaborations with other leading centers to ensure that NSF's vision of petascale science is achieved through the Track2/Track1 deployments.

• 3/28/2007 - Talk - April Lorenzen
  April Lorenzen talk in SDSC Room 408 at 1:00 P.M. on Wednesday, March 28, 2007.
  
  

  Abstract:

  In my talk I will describe the data I have been collecting since 2004: every gTLD name server host, TLD registry change of gTLD name server host IP, every appearance / disappearance / move of a gTLD domain from one name server host to another. I did this every time each new gTLD zone file is released by its registry - 1x to 2x daily.

  Anyone can get an account and download a TLD zone file, but I've added a time component. The zone files are too big for most people to keep around more than the latest copy. I loaded every daily snapshot into into postgres, timestamped it, and created mapping tables that relate the data in a multitude of extremely useful ways, especially for tracking spam.

  I would also like to address why an individual like myself was the one to do this, and where this data should live and who should have access to it, since there are commercial as well as public sector applications.

[Back to Calendar View]

April

• 4/13/2007 - Talk - Steve Cutchin
  Steve Cutchin (SDSC) talk "Visualization at SDSC using TeraGrid" in SDSC Room 462 from 11:00 AM - 12:30 AM. Part of the Cyberinfrastructure Seminar Series. The talk will also be webcast at http://www.cichannel.org
  
  

  SDSC visualization services creates animations, images, and customized tools for users and researchers associated with the San Diego Supercomputer Center. To enable our capability to provide visualization support to a wide range of users we have developed a collection of tools, portals, and services that allow us to do compelling visualization on behalf of our users utilizing Teragrid Resources both locally and remote. In this talk I will present an overview of these capabilities as well as showcase specific visualizations created using them.

  The Cyberinfrastructure Seminar Series is a set of presentations on cyberinfrastructure and related research organized by NCSA and SDSC. All Access Grid sites are welcome to participate in this seminar. For questions regarding this event, contact the SDSC Training Group at training-coord@sdsc.edu.

• 4/19/2007 - Talk - Vinton Cerf
  Vinton Cerf (Google) talk "Tracking the Internet into the 21st Century, Part Two" in the main CalIT2 Auditorium in Atkinson Hall at UCSD. This talk will also be webcast at http://live-calit2.ucsd.edu:8080/ramgen/broadcast/live.rm
  
  

  DESCRIPTION/ABSTRACT:
  The indomitable minds at Google see opportunities that are a consequence of the rapid flow of information into the Internet and the expanding infrastructure that supports Internet access and use. The merging of media, and the transformation of traditional media through download and playback, is changing the face of conventional advertising; likewise, being online as advertising information becomes available changes the way in which consumers interact with the data. And, of course, consumers are becoming producers as evidenced by the YouTube and blogging phenomena. Mobility is adding into the mix, as is geographically indexed information. Billions of devices on the Internet, reachable and controllable through it, add another twist to the opportunity space. Hear first-hand Google's predictions on the changing face of conventional advertising and the ever-expanding infrastructure that supports Internet access and use. See how the $130-billion company is sizing up new opportunities with current market trends; from the merging of traditional and online media to the consequences of interactive and user-generated content.

  SPEAKER BIO:
  Vint Cerf is a member of Calit2's Advisory Board and is known as a "Father of the Internet" for his invention of the TCP/IP protocol and his work on Internet architecture. He served at DARPA as it developed TCP/IP packet switching networks and the ARPANET, he developed at MCI the first commercial email service to be connected to the Internet, and he joined Google in September of 2005 as its Chief Internet Evangelist (yes, that's his official title). For his contributions, he has won almost every award in the Internet/Communications field, including in 1997 the U.S. National Medal of Technology; and in 2005, the highest civilian honor bestowed in the U.S., the Presidential Medal of Freedom, in recognition of his work on the software code used to transmit data across the Internet that "enabled the digital revolution and transformed global commerce, communication, and entertainment." Cerf did undergraduate and graduate school at Stanford in mathematics, before doing his Ph.D. in computer science at UCLA.

[Back to Calendar View]

May

• 5/4/2007 - Talk - Dr. Donald E. Brownlee
  Dr. Donald E. Brownlee gives a talk in the James R. Arnold lecture series entitled The STARDUST Mission in the Natural Sciences Building auditorium (at UCSD in Revelle College) on Friday, May 4, 2007 at 4:00 PM.
  
  

  An endowed lectureship has been established to honor Professor James R. Arnold, one of UCSD's first faculty members and the founding chair of the Chemistry Department. The sixth James R. Arnold Lecture, will be given by Dr. Donald E. Brownlee, Professor of Astronomy at the University of Washington. Dr. Brownlee will share his exciting findings from the NASA STARDUST mission to collect particles from Comet Wild 2. The event will be held at the University of California, San Diego (UCSD) campus at Revelle College in the Natural Sciences Building Auditorium at 4:00PM on Friday, May 4th.

  This lecture is free and open to the public. UCSD Parking permits are required to park on campus, you can purchase a permit for $6.00 for the evening. For more information please see the web page: http://calspace.ucsd.edu/casgc/JArnoldLecture/Home.html or contact Tehseen Lazzouni at (858) 822-1597 or tlazzouni@ucsd.edu .

  ABSTRACT
  The NASA STARDUST mission collected thousands of particles from Comet Wild 2 that are now being studied by two hundred scientists around the world. The spacecraft captured the samples during a close flyby of the comet in 2004 and returned them to Earth with a dramatic entry into the atmosphere early in 2006. The precious cargo of comet dust is being studied to determine new information about the origin of the Sun and planets. The comet formed at the edge of the solar system, beyond the orbit of Neptune, and is a sample of the material from which the solar system was formed. One of the most dramatic early findings from the mission was that a comet that formed in the coldest place in the solar system contained minerals that formed in the hottest place in the solar system. The comet samples are telling stories of fire and ice and they providing fascinating and unexpected information about our origins.

  SPEAKER BIO
  Don Brownlee is a Professor of Astronomy at the University of Washington. His research interests include interplanetary dust, meteorites, comets, the origin of the solar system, astrobiology, and STARDUST?a NASA Discovery mission to collect samples from a comet and bring them back to Earth. Don began collecting and analyzing space samples as a graduate student, when he conducted high altitude balloon flights to collect extraterrestrial material and built micrometeoroid collectors that were flown on NASA missions. Don has also studied the size distribution of impact craters on spacecraft surfaces and conducted chemical analysis of meteoroid residue inside craters.

  Don and Dr. Peter Tsou of Jet Propulsion Laboratory devised a plan to collect samples from Halley?s comet prior to 1986 and this was the seed for the development of the successful NASA Stardust mission that returned comet samples to Earth on January 15, 2006. To further develop collection techniques they flew aerogel on the European Recoverable Carrier (Eureca) which successfully collected particles and verified the use of aerogel as a space collector.

  Don received his B.S. in Electrical Engineering at the University of California, Berkeley in 1965 and his Ph.D. in Astronomy from the University of Washington in 1971. Don has received numerous awards and honors including the NASA Medal for Exceptional Scientific Achievement, Asteroid Namesake 3259 Brownlee, the J. Lawrence Smith medal from the National Academy of Sciences, the Leonard medal from the Meteoritical Society, the Popular Mechanics?Breakthrough Award (STARDUST), the Aviation Week?Program Excellence Award (STARDUST) and the Aviation Week 2007 Laureate Award for Space. Don has been elected fellow of the American Astronomical Society, American Association for the Advancement of Science, the Meteoritical Society, and the American Geophysical Union. Don is currently an Associate Editor of Meteoritics, a Principal Investigator for STARDUST, and a Science Working Group Member of Kepler, a NASA Discovery mission to find habitable planets, scheduled to launch in 2008.

  Don is the co-author of two books: "The Life and Death of Planet Earth" and "Rare Earth: Why Complex Life Is Uncommon in the Universe?".

  His website is: http://www.astro.washington.edu/brownlee.

• 5/14/2007 - Talk - Larry Peterson
  Larry Peterson (Princeton University) gives a talk entitled PlanetLab: Evolution vs Intelligent Design in Global Network Infrastructure on Monday, May 14 2007 from 1:00 PM - 2:00 PM in EBU3B 1202.
  
  

  Abstract:
  PlanetLab is a global platform for evaluating and deploying network services. It currently includes over 750 nodes, spanning nearly 380 sites and 30 countries, and hosts over 600 experimental services. PlanetLab must satisfy a unique set of sometimes contradictory requirements: based on our experiences building PlanetLab over the past three years, we are now able to define an architecture that satisfies these requirements. This talk identifies the requirements, presents the design principles that follow from them, and outlines the resulting PlanetLab architecture. It also briefly discusses some of the lessons we learned about building large network systems.

  Speaker Bio:
  Larry Peterson is the Robert E. Kahn Professor of Computer Science at Princeton University. He is Director of the Princeton-hosted PlanetLab Consortiu