CSE 600 (Ongoing Research Seminar) Fall 2005

Abstract:

Software bugs cause almost all security vulnerabilities, and attackers exploit these weaknesses to send Spam, launch DDOS attacks, and defraud online banks and merchants. By making improvements in software security, we can prevent many of these attacks. I will present new static analysis techniques, based on type qualifier inference, for automatically and efficiently finding bugs in large software systems. This approach has many advantages, including scalability, precision, usability, and soundness, enabling it to verify the absence of bugs, which is critical to security. I have implemented these methods in a static analysis tool, CQual, and used it to find dozens of bugs in the Linux kernel.

Abstract:

Location-based protocols such as geographical forwarding have been considered as efficient and scalable schemes for large wireless ad hoc networks. However, for a wide variety of sensor network environments, location information is unavailable or expensive to obtain. In the talk I will present location-free routing schemes in sensor networks. The main intuition is to explore the underlying global topology of the sensor field and construct virtual coordinates for sensor nodes that enable local and efficient gradient-based routing. I will talk about two schemes in this family: Gradient Landmark-based Distributed Routing protocol (GLIDER), and Medial Axis-based Geometric Routing protocol (MAP). Both are two-level routing structures where sensor nodes are partitioned into tiles with nice geometry and connectivity so that localized and efficient routing can be achieved inside a tile. We use global topological information to glue the routable tiles together to enable efficient cross-tile routing. Such topology-based schemes use no geographic information, make few assumptions on the network model, and achieve good load balancing. They compare favorably with previously proposed geographical schemes.

Abstract:

In this talk I am going to chat about a bunch of things related to security and why it is fun to work in this area. I will talk about established research and then about some new stuff i am working on. The talk will (briefly) touch on Quantum Cryptography, evil Mallory, innocent Alice, (not so) innocent Bob, naive Caroll (only sometimes), beautiful Eve and trusted Trent. Then we discuss briefly about secure hardware and (maybe) biometrics -- as we are about to purchase a bunch of PDAs, GPS devices and iris scanners and play around with them for (mainly) security reasons. If time will permit i will also pitch in my 5c about academia and a PhD. The talk should last about 1 hour. Feel free to bring your significant other or a good book in case of a boredom attack. See you all there.

Abstract:

Surface appearance is one of the most important aspects of visual realism of computer graphics imagery. This talk will provide an overview of several modern approaches to treating some aspects of this problem, concentrating on direct measurement of surface reflection properties and recently developed analytical microfacet-based reflection models.

Abstract:

Cellular networks are a critical component of the economic and social infrastructures in which we live. In addition to voice services, these networks deliver alphanumeric text messages to the vast majority of wireless subscribers. To encourage the expansion of this new service, telecommunications companies offer connections between their networks and the Internet. The ramifications of such connections, however, have not been fully recognized. In this paper, we evaluate the security impact of the SMS interface on the availability of the cellular phone network. Specifically, we demonstrate the ability to deny voice service to cities the size of Washington D.C. and Manhattan with little more than a cable modem. Moreover, attacks targeting the entire United States are feasible with resources available to medium-sized zombie networks. This analysis begins with an exploration of the structure of cellular networks. We then characterize network behavior and explore a number of reconnaissance techniques aimed at effectively targeting attacks on these systems. We conclude by discussing countermeasures that mitigate or eliminate the threats introduced by these attacks.

Speaker Bio:

Patrick McDaniel is the Hartz Family Career Development Professor in the Computer Science and Engineering Department at the Pennsylvania State University, and director of the Systems and Internet Infrastructure Security Laboratory. He received his Ph.D. from the University of Michigan in 2001 where he studied the form, algorithmic limits, and enforcement of security policy. Prior to joining Penn State, Patrick was a senior technical staff Member of the Secure Systems Group at AT&T Labs-Research and Adjunct Professor of the Stern School of Business at New York University. Patrick's recent research efforts have focused on security management in distributed systems, network security, and public policy and technical issues in digital media. Patrick is a past recipient of the NASA Kennedy Space Center fellowship, a frequent contributor to the IETF security standards, and has authored many papers and book chapters in various areas of systems security. He is currently serving as the the Program Chair of the 2005 USENIX Security Symposium, the Vice Chair for Security and Privacy for WWW 2005, and is the Chair of the Industry and Government Track at the ACM Computer and Communications Security conference. Patrick is also an associate editor of the journal ACM Transactions on Internet Technologies. Prior to pursuing his Ph.D. in 1996, Patrick was a software architect and program manager in the telecommunications industry.

Abstract:

Information integration consists in offering a uniform access to a set of heterogeneous data sources. A user of an information integration system accesses the underlying data through a uniform representation called global view; when she/he issues a query, the system issues in turn suitable queries to the different sources of data, and assembles the results. Since the global view represents the domain of interests, it is important that it is expressed with a flexible and expressive formalism; therefore, the global schema is enriched with constraints. In this talk we show techniques and algorithms for information integration systems, based on the relational data model, and with different classes of constraints on the global schema. We show that the presence of constraints significantly complicates the problem, since it forces the system to reason on incomplete information. However, tractable algorithms exist for most practical cases.

Abstract:

Benchmarking is critical when evaluating the performance of new file system code, but doing so correctly is difficult. Whereas care must be taken when benchmarking any system, this is especially true with file systems, for two reasons. (1) Complex interactions between I/O devices, specialized caches, kernel daemons, and other operating system components result in behavior that is difficult to analyze. (2) each file system has its own features and is optimized for certain conditions and workloads, so there is no single, uniform way to benchmark every file system. We have found that some of the most commonly used benchmarks are flawed, and many research papers do not provide a clear enough picture of file system performance. We believe that a good performance evaluation should use micro-benchmarks to highlight both the good and bad qualities of a file system, as well as general-purpose benchmarks or traces to give an idea about how it would perform under expected and realistic workloads. Nevertheless, care should be taken to ensure that general-purpose benchmarks indeed accurately reflect the real-world workloads. In addition, benchmarks should scale well, and results should be reproducible and comparable across papers. In this talk, we first present a survey of file system benchmarks used in 68 recent research papers. We found that no single benchmark adequately measures file system performance. We then show how some commonly acceptable and widely used benchmarks and benchmarking techniques can easily conceal overheads, unfairly over-emphasize overheads, or can in general emphasize or de-emphasize many of the file system's properties. We then present suggestions on how to create and conduct benchmarks so that they provide a more fair and accurate picture of file system performance. We end the talk by describing our views on the future of file system benchmarking. To that end, we have been developing several technologies that will be described in this talk: fine-grained file system tracing, efficient file system replaying, automated file system benchmarking tools, and low-overhead detailed file system behavior visualization tools.

Speaker Bio:

EREZ ZADOK, Assistant Professor
Stony Brook University, Computer Science Department

Ph.D. 2001, Columbia University

Erez Zadok's research focuses on operating systems, with a specialty in file systems and storage. He studies operating systems and file systems from many aspects: security, efficiency, scalability, reliability, portability, survivability, usability, ease-of-use, versatility, flexibility, and more. Special attention is given to balancing three often-conflicting aspects of computer systems: security, performance, and ease-of-use. Since joining Stony Brook in 2001, Zadok and his group in the Filesystems and Storage Lab (FSL) developed many file systems and operating system extensions; examples include a highly-secure cryptographic file system, a portable versioning file system, a tracing file system useful to detect intrusions, a snapshotting and sandboxing file system, an anti-virus file system, an integrity-checking file system, a compiler to convert user-level C code to in-kernel efficient yet safe code, stackable file system templates, and more. Zadok's research is supported by several NSF grants including an NSF CAREER award, an NSF Trusted Computing grant, an NSF CSR award, and two joint NSF awards for Information Assurance Education (Capacity Building and Scholarship for Service). Zadok's lab exposes students to internals of over a dozen different operating systems. Zadok is the author of "Linux NFS and Automounter Administration" (Sybex, 2001).