EMİN GÜN SİRER
	Associate Professor 4151 Upson Hall Computer Science Dept. Cornell University Ithaca, NY 14853	(607) 255-7673 (607) 255-4428 fax Turn on JavaScript to view email address public key

Talks

• High-Performance Peer-to-Peer Overlays, or, Heuristics Considered Harmful.
  University of Michigan, Ann Arbor, MI, June 2006.
  Invited Colloquium: Rennselaer Polytechnic Institue, Troy, NY, May 2006.
  
  In this talk, I will describe a new approach for building distributed systems with strong performance and resource usage guarantees. The critical insight behind this work is to formalize the core tradeoffs in distributed systems as a mathematical optimization problem. We can then achieve high performance in the presence of limited resources by minimizing a targeted performance function subject to constraints in a distributed fashion.

  The talk will outline this new approach and describe how we recently applied it to build three peer-to-peer systems: CoDoNS, a replacement for DNS, CobWeb, an open-access content distribution network like Akamai, and Corona, an RSS-like system for disseminating Web micronews. All three systems have been deployed on PlanetLab, and either guarantee near-optimal lookup/update performance subject to bandwidth constraints or achieve a targeted level of lookup/update performance while minimizing bandwidth and storage costs. Overall, this approach represents a novel way to building large-scale distributed systems that contrasts with past systems based on ad hoc heuristics.

• Perils of Transitive Trust, or How to 0wn the Internet via DNS.
  Invited Plenary Talk: RIPE-52, Istanbul, Turkey, April 2006.
  
  DNS delegations create a complex web of trust with subtle dependences that make it all too easy to hijack domains. Our recent survey found that attackers exploiting well-documented vulnerabilities in DNS nameservers can hijack more than 30% of the names appearing in the Yahoo and DMOZ directories, more than 84% when coupled with modest-sized, well-timed DoS attacks.

• Trustworthy Computing with the Nexus Operating System.
  UC Berkeley, Berkeley, CA, April 2006.
  CMU, Pittsburgh, PA, January 2006.
  
  Tamper-proof coprocessors for secure computing are poised to become a standard hardware feature on future computers. Such hardware provides the primitives necessary to support trustworthy computing applications, that is, applications that can provide strong guarantees about their run time behavior. Current operating systems, however, lack the architecture and abstractions required to support trustworthy computing.

  In this talk, I will outline a new operating system, called the Nexus, my group is building to support trustworthy computing applications. The Nexus has three unique properties: a novel architecture to reduce the trusted computing base, a general-purpose and flexible attestation mechanism to establish statements about the current state of a computation, and a strong, high-performance isolation mechanism to enable reasoning about future behavior based on statements about the present. This talk will describe the new abstractions and mechanisms we developed to support trustworthy computing, illustrate their use with several applications we have built, and point to new research directions enabled by a trustworthy operating system.

• High Performance Peer-to-Peer Systems, or Heuristics Considered Harmful.
  University of Toronto, Toronto, Canada, January 2006.
  
  In this talk, I will describe a new approach for building distributed systems with strong performance and resource usage guarantees. The critical insight behind this work is to formalize the core tradeoffs in distributed systems as a mathematical optimization problem. We can then achieve high performance in the presence of limited resources by minimizing a targeted performance function subject to constraints in a distributed fashion.

  The talk will outline this new approach and describe how we recently applied it to build three peer-to-peer systems: CoDoNS, a replacement for DNS, CobWeb, an open-access content distribution network like Akamai, and Corona, an RSS-like system for disseminating Web micronews. All three systems have been deployed on PlanetLab, and either guarantee near-optimal lookup/update performance subject to bandwidth constraints or achieve a targeted level of lookup/update performance while minimizing bandwidth and storage costs. Overall, this approach represents a novel way to building large-scale distributed systems that contrasts with past systems based on ad hoc heuristics.

• Peer-to-Peer: Still Useless?.
  Panel Member, Symposium on Operating Principles, Brighton, UK, October 2005.
  
  I believe peer-to-peer systems offer a drastically new and very promising way to build robust infrastructure services. The area, coupled with the success of PlanetLab, has revitalized the distributed systems area, opened up many new avenues for research, already delivered interesting new systems and there is so much more to come.

• Network Positioning for Wide-Area and Wireless Networks..
  Keynote Talk: LOCALITY Workshop, Cracow, Poland, September 2005.
  
  Determining the location of nodes in a network is a basic building block operation that enables many interesting, location-aware applications. In this talk, I will describe two techniques for determining node positions: the first is a comprehensive localization system for wireless networks, and the second is a localization and routing framework for Internet-like wide-area networks.

  In wireless networks, I believe the challenge is to efficiently pinpoint the location of nodes without having to use expensive and energy- consuming specialized hardware such as GPS receivers. I will describe a new framework for wireless localization in which we cast the problem as a constraint system, extract constraints aggressively from the MAC layer, and solve it with the aid of a few landmarks. Simulations and deployment on Berkeley motes and laptops indicate that this approach achieves high accuracy.

  In wide-area networks, I believe the challenge is to efficiently resolve geographic queries (e.g. "where is the closest node to www.cnn.com?") at high scale. Previous approaches relied on network coordinates; in this talk, I will outline a scalable and efficient scheme for resolving geographic queries without computing network coordinates, and show how this framework can be applied to three qualitatively different wide-area applications.

• Issues in Dependability for Self-Organizing Nomadic Systems.
  Invited Talk: IFIP 10.4 Workgroup Meeting, Hakone, Japan, July 2005.
  

• Self-Organizing Systems for Robust, Large-Scale Infrastructure Services.
  University of Maryland, College Park, Maryland, July 2005.
  
  In this talk, I will present some recent work on building new, robust infrastructure services based on self-organizing, peer-to-peer architectures. In particular, I will describe Herbivore, a system based on Dining Cryptographer Networks for anonymous communication that provides strong anonymity, scalability, and performance. In addition, I will touch upon some recent work in the Honeycomb project, where we are building new infrastructure applications, such as a replacement for DNS, a new CDN and a new pub-sub system, that provide high-performance in the presence of highly-skewed distributions, traditionally considered to pose difficult problems for networked services.

• Robust and Attack-Resilient Self-Organizing Networks.
  Workshop on the Complex Behavior of Adaptive Network-Centric Systems, College Park, Maryland, July 2005.
  

• A Comprehensive Framework for Dependable Nomadic Computing.
  Invited Speaker: Workshop on Dependable Nomadic Computing, IFIP Working Group on Dependable Computing 10.4, Hakone, Japan, July 2005.
  

• Trustworthy Computing at Cornell Computer Science.
  TRUST Retreat, Santa Cruz, California, June 2005.
  

• Some Challenges in Building Internet-Scale Distributed Services.
  Invited Speaker: Cornell University ACSU, April 2005.
  
  Why I love doing what I do.

• CoDoNS: A High-Performance, Fault-Resilient Replacement for the Legacy Domain Name Service.
  Carnegie Mellon University, Computer Science Department, November 2004.
  
  The Domain Name Service (DNS) is based on a static, manually constructed hierarchy of distributed nameservers. This organization makes the legacy DNS slow, vulnerable to denial-of-service attacks, and prone to load imbalance.

  In this talk, I will describe an alternative architecture for a new Domain Name Service that provides high performance, DoS-resilience, and automatic load balancing in the presence of flash crowds and the "slashdot effect." This talk will focus on the core contribution, proactive caching, which enables scalable peer-to-peer overlays to respond to queries drawn from a Zipf-distribution in under a single hop with a minimal number of object replicas, and illustrate its application to DNS. The resulting system, CoDoNS, has been deployed throughout the globe over Planetlab and can serve either as a high-performance replacement for legacy DNS or as a backup in case legacy DNS comes under attack.

• Graduate School: The End Game.
  CAREER talk, Cornell, October 2004.
  
  Advice to graduating PhD students on how to get a permanent research position.

  In this talk, I will discuss the process of finishing graduate school and finding a job, with special emphasis on interviewing for academic research positions. I'll talk about the timeline, the required preparation, the job talk, the interviews, and the offers. While the talk is intended particularly for students close to graduation, students who plan to graduate at some point should attend it in order to plan ahead for the final phases of graduate school.
  (Cornell-only)

• Eluding Carnivores: File Sharing with Strong Anonymity.
  European SIGOPS Workshop, Leuven, Belgium, September 2004.
  
  Anonymity is increasingly important for networked applications amidst concerns over censorship and privacy. This paper outlines the design of a scalable and efficient file sharing system that provides strong anonymity. Herbivore provides k-anonymity against powerful adversaries able to monitor and manipulate all network traffic. Herbivore achieves scalability by partitioning the global network into smaller anonymizing cliques. Measurements on PlanetLab indicate that the system achieves high anonymous bandwidth when deployed on the Internet.

• Self-Organizing Networks.
  NSF STC: TRUST Center, University of California, Berkeley, September 2004.
  
  Self-organization is a required property for the types of large-scale, complex networked systems we are likely to encounter in the future. In this presentation to the NSF site review panel, I outline MagnetOS, a self-organizing operating system for ad hoc and sensor networks, and CoDoNS, a self-organizing name service for wide area networks.

• The Design and Implementation of a Next Generation Name Service for the Internet.
  SIGCOMM, Portland, Oregon, August 2004.
  
  CoDoNS is a safety net and replacement for the legacy DNS system. It provides high performance by intelligently utilizing bandwidth and storage to replicate popular objects on top of a peer-to-peer, self-organizing overlay.

• CoDoNS: A Next Generation Name Service.
  DNSSEC Steering Committee, Washington, DC, August 2004.
  
  In this talk, I describe CoDoNS, and how it can help speed up the adoption of DNSSEC. DNSSEC has been still-born; few people use it, despite its obvious advantages. The US government is busy deploying DNSSEC-enables nameservers for names in its jurisdiction. My thesis is that CoDoNS, which attaches DNSSEC signatures to names obtained via legacy DNS, can help provide signatures even for those domains that were not signed in the first place. This enables clients to enjoy the benefits of secure, tamperproof name service, and provides a stepping stone towards wider adoption of DNSSEC.

• Operating Services 7/24: Domain Name Service for the Next Generation Internet.
  Planetlab Workshop, Palo Alto, California, April 2004.
  
  In this talk, I provide an overview of the CoDoNS system, how it is deployed on PlanetLab, and how it can drastically change the way we think of and implement name services by liberating names from the servers that provide them. This breaks the monopoly that namespace operators enjoy over top level domains, and makes possible a new kind of service.

• CoDoNS: A High-Performance, Fault-Resilient Replacement for the Legacy Domain Name Service.
  University of Pennsylvania, March 2004.
  Microsoft Research Silicon Valley, March 2004.
  
  The Domain Name Service (DNS) is based on a static, manually constructed hierarchy of distributed nameservers. This organization makes the legacy DNS slow, vulnerable to denial-of-service attacks, and prone to load imbalance.

  In this talk, I will describe an alternative architecture for a new Domain Name Service that provides high performance, DoS-resilience, and automatic load balancing in the presence of flash crowds and the "slashdot effect." This talk will focus on the core contribution, proactive caching, which enables scalable peer-to-peer overlays to respond to queries drawn from a Zipf-distribution in under a single hop with a minimal number of object replicas, and illustrate its application to DNS. The resulting system, CoDoNS, has been deployed throughout the globe over Planetlab and can serve either as a high-performance replacement for legacy DNS or as a backup in case legacy DNS comes under attack.

• Peer-to-Peer Authentication With a Distributed Single Sign-On Service.
  International Workshop on Peer-to-Peer Systems, San Diego, California, February 2004.
  
  The right (fault-tolerant, scalable, flexible) way to do what Microsoft's passport.com set out to do. CorSSO creates an open market for authentication -- any peer can serve as an authenticator, and compete with other peers on price, performance, and security.

• KARMA: A Secure Economic Framework for P2P Resource Sharing.
  Workshop on Economics of Peer-to-Peer Systems, Berkeley, California, Spring 2003.
  
  Peer-to-peer systems are typically designed around the assumption that all peers will willingly contribute resources to a global pool. They thus suffer from freeloaders, that is, participants who consume many more resources than they contribute. In this talk, I outline how the KARMA system provides a general economic framework for avoiding freeloaders in peer-to-peer systems. KARMA works by keeping track of the resource consumption and resource contribution of each participant. The overall standing of each participant in the system is represented by a single scalar value, called their karma. A set of nodes, called a bankset, keeps track of each node's karma, increasing it for resource contributions and decreasing it for consumption. KARMA is resistant to malicious attacks by the provider, consumer and a fraction of the bankset.

• Doing Research in a Group Setting.
  CAREER talk, Cornell University, Spring 2003.
  

• Operating System Support for Ad Hoc Networks in MagnetOS.
  University of Rochester Colloqium Series, Rochester, NY, September 30, 2002.
  
  An overview of the MagnetOS distributed operating system for ad hoc networks.

• NSF Research Infrastructure Grant: A3 - Anytime, Anywhere, Anything.
  NSF PI Meeting, Snowbird, Utah, Summer 2002.
  

• An Access Control Language for Web Services.
  ACM Symposium on Access Control Models and Technologies (SACMAT), Naval Postgraduate School, Monterey, California, June 3-4, 2002.
  
  An overview of the web reference monitor we have developed as part of the WebGuard project. WebGuard factors safety out of web service implementations into a separable high-performance module.

• Issues in Ubiquitous Computing.
  Invited Panelist: International Performance Computing and Communications Conference, Phoenix, Arizona, April 2002.
  
  One of the primary emerging challenges for ubiquitous computing is how to manage the hundreds of devices each individual will soon possess and interact with. My group is developing a new operating system that makes an entire collection of such devices act like a single virtual machine, greatly simplifying the management and programmability of the network.

• Operating Systems Support for Ad Hoc Networks.
  Cornell Brown Bag Lunch, Spring 2002.
  

• MagnetOS: An Operating System for Ad Hoc Networks.
  Symposium on Operating Systems WIP, Lake Louise, Banff, Canada, October 2001.
  
  In this WIP talk, I outline the new operating system we are building for wireless ad hoc and sensor networks. Making the entire network appear as a single virtual machine expands the freedom available to the system for power-efficient execution and simplifies application programming.

• Securing System Code.
  EGSA Invited Talk, Ithaca, New York, Arpil 2001.
  

• SPIN: A Retrospective on an Extensible System.
  Cornell Brown Bag Lunch, March 2001.
  
  SPIN was one of the few new operating systems developed in academia in the 90's that provided a novel extensibility mechanism based on typesafe languages. I designed and implemented most of the SPIN kernel as well as various extensions. In this talk, I will provide a brief overview of SPIN, and describe what went well, what did not, and what the remaining challenges are.

• Distributed Virtual Machines: A New System Architecture for Networked Computers.
  University of Pennsylvania, March 2000.
  University of Colorado, March 2000.
  Compaq Systems Research Center, March 2000.
  Dartmouth College, March 2000.
  University of Utah, April 2000.
  Oregon Graduate Institue, April 2000.
  Microsoft Research, April 2000.
  Massachusetts Institute of Technology, April 2000.
  Yale University, April 2000.
  Cornell University, May 2000.
  
  The rise of the Internet, the explosion in the number of networked hosts, and the accompanying shift from desktop computing on workstations to network computing on ubiquitous devices, have outpaced the capabilities of modern operating systems. In particular, even though the Java virtual machine was specifically designed to offer strong security guarantees, to simplify system management, and to support cheap embedded devices, Java installations have had difficulty achieving these goals.

  In this talk, I will describe a new system architecture that addresses the manageability, security and scalability requirements of networked computers. This architecture factors operating system services out of clients and locates them on transparent proxies within the networking fabric. I will describe our distributed Java virtual machine implementation based on this architecture, and show that this approach can yield smaller, cheaper clients and more manageable and more secure networks.

• Design and Implementation of a Distributed Virtual Machine for Networked Computers.
  Symposium on Operating Systems Principles, Kiawah Island, South Carolina, December 1999.
  
  Describes the factored service architecture of the Kimera Project, and illustrates how such a design can lead to manageable, secure and scalable implementations of traditional virtual machine services, as well as provide a platform for new services.

• Testing Large Systems Using Production Grammars.
  Microsoft Research, Redmond, Washington, November 1999.
  
  Expanded version of the DSL talk below, describing how a tiny team could build a robust Java virtual machine using automated testing techniques.

• Testing Java Virtual Machines.
  International Conference on Software Testing And Review, San Jose, California, November 1999.
  
  Describes three testing techniques, namely random variations, production grammars and extended production grammars with certificates, and relates our experience with using them to create robust Java virtual machines.

• Using Production Grammars in Software Testing.
  Second Symposium on Domain Specific Languages, Austin, Texas, October 1999.
  
  Description of the domain-specific language lava that we developed for specifying production grammars. Also describes how the language was used to generate comprehensive test suites for the Java virtual machine.

• A Practical Approach for Improving Startup Latency for Java Applications.
  Workshop on Compiler Support for Systems Software, Atlanta, Georgia, May 1999.
  
  This talk describes a technique to restructure Java applications such that their startup times are reduced.

• Interaction of I/O and Memory Subsystem Design with Operating Systems.
  CSE 548: Advanced Computer Architecture. Invited Lecture, April 1999.
  
  Four lectures on the interaction of architectural design at the I/O and memory subsystem level with operating system functionality, given in a graduate-level class on architecture.

• Distributed Virtual Machines: A System Architecture for Network Computing.
  Eighth ACM SIGOPS European Workshop, Sintra, Portugal, September 1998.
  
  Covers three different techniques for migrating system services out of network end-points into servers, and describes why we chose binary rewriting in the Kimera project.

• Networks and Operating Systems for Smart Spaces.
  DARPA Graduate Student Workshop, Rosslyn, Virginia, July 1998.
  
  The results of the operating systems workgroup at the DARPA workshop. The talk discusses operating system techniques for supporting smart-space applications, the workshop theme.

• A System Architecture for Next Generation Network Computing.
  DARPA Graduate Student Workshop, Rosslyn, Virginia, July 1998.
  
  Covers the goals, vision and accomplishments of the Kimera project for our funding agency.

• Extensibility, Safety and Performance in the SPIN Operating System.
  CSE 590YA: Operating Systems. Invited Lecture, February 1998.
  
  A description of the SPIN operating system, given in a masters-level class on operating systems.

• Verifying Verifiers: Techniques for Building Robust Virtual Machines.
  Intel Research Labs, Portland, Oregon, February 1998.
  
  Overview of the research topics covered by the Kimera project.

• Verifying Java Verifiers.
  Workshop on Security and Languages. Palo Alto, California, October 1997.
  
  Description of a simple and highly effective technique for automatically creating test cases for Java virtual machines. Brief discussion of the types of security flaws uncovered using this technique.

• Protection in Operating Systems.
  CSE 451: Operating Systems. Invited Lecture, October 1997.
  
  Description of a language-based technique for creating multiple protection domains within a single, protected hardware address space, given in the undergraduate class on operating systems.

• Project Kimera.
  Vortex/Diesel Project Retreat. Invited Speaker, Blaine, Washington, June 1997.
  
  Overview of the Kimera project. A proposal to explore language techniques for automatically eliminating unnecessary synchronization operations, a common source of inefficiency in typical Java applications.

• Security in the Java Virtual Machine.
  Seattle Java User's Group. Invited Speaker. Seattle, Washington, June 1997.
  
  Description of the Kimera project's goals and approach. Brief discussion of flaws uncovered in commercial virtual machines.

• Extensibility, Safety and Performance in the SPIN Operating System.
  CSE 551: Advanced Operating Systems. Guest Lecture, November 1996.
  
  Overview of the SPIN operating system, given in a graduate-level class on operating systems.

• Extensible Routing.
  SPIN/Loom Retreat. Kalaloch Lodge, Olympic Peninsula, Washington, October 1996.
  
  A proposal to use an extensible system, such as SPIN, over a cluster to implement an extensible router. In 1993, when I first started working building the SPIN kernel, I was planning to build up the kernel and do my thesis on a router whose processing code could be modified dynamically. David Tennenhouse publicly mentioned the idea at OSDI, coined the term Active Networking and scooped the idea. This talk describes research directions we can pursue based on the SPIN infrastructure that the Active Networking community is missing.

• Replicated Web Search Engines.
  SPIN/Loom Retreat. Kalaloch Lodge, Olympic Peninsula, Washington, October 1996.
  
  Internal architecture of some highly scalable search engines. Big iron, SMP approach of AltaVista versus PC-cluster approach in Inktomi.

• Synchronization Alternatives for Clusters of Workstations.
  SPIN/Loom Retreat. Kalaloch Lodge, Olympic Peninsula, Washington, October 1996.
  

• Consensus.
  SPIN/Loom Retreat. Kalaloch Lodge, Olympic Peninsula, Washington, October 1996.
  
  Survey of techniques for achieving consensus in a tightly connected cluster and a detailed description of the Paxos quorum algorithm.

• Language Support for Extensible Operating Systems.
  First Annual Workshop on Compiler Support for System Software, Tucson, Arizona, February 1996.
  
  A talk that covers three papers on language support for operating systems, extensions to Modula-3 to support demanding applications, and a dynamic-linking based technique for establishing and manipulating protection domains within a single protected address space.