Keeping up-to-date with current research is a critical task for both students and scholars. A weekly seminar is a fun and social way to keep in touch with other's work. (Doing it by yourself is boring!) At the seminar, you can eat a few cookies, chitchat about the finer points of finer points of mutual exclusion, and exchange ideas with students and faculty working in your field.
Our mailing list is os-seminar@cs.wisc.edu. To subscribe, send mail to majordomo@cs.wisc.edu with subscribe os-seminar in the body. The list traffic is about one message per week to announce the next seminar.
Questions about the seminar and arrangements may be directed to Douglas Thain
Date | Location | Topic and Speaker | Related Material
Friday, | 15 September 2:30-3:30pm Room 1221
| Meet the Faculty | This seminar will be a fun poster session! After a brief introduction, faculty and students will present posters of current systems research. You can mill around, learn what's going on, and most importantly, meet people and ask questions. Snacks will be provided!
| Monday, | 18 September 4-5 pm Room 2310
| Interposition Agents for Distributed Computing | Douglas Thain Many distributed systems provide novel features that users are eager to take advantage of. However, each new system requires programmers willing to rewrite their applications. Rewriting programs for every potential new distributed system is tiresome and non-scalable. I will demonstrate how interposition agents can be used to attach legacy programs to new systems and comment on the problems that result.
slides | web page Monday, | 2 October 4-5 pm Room 2310
| Fine-Grained Dynamic Kernel Instrumentation | Ariel Tamches This talk will present an overview of the KernInst group's research on runtime operating system instrumentation. Viewed from a low level, we have designed and implemented a technology that allows runtime-generated code to be inserted into a Solaris/UltraSparc kernel. Of special note is that (1) instrumentation takes place on a stock, off-the-shelf, already running kernel, (2) instrumentation is fine-grained; code can be inserted before arbitrary machine code instructions, and (3) instrumentation takes place entirely at runtime, without any need to recompile, reboot, or even pause the kernel. This talk will also give an overview of several interesting tools (some completed; some still in development) that use fine-grained kernel instrumentation. One, kperfmon, allows a user to collect detailed performance metrics (such as elapsed time, virtual time, and time spent processing cache misses) on kernel functions. A second, kcc, measures fine-grained kernel code coverage "for free" (i.e., with almost zero runtime overhead). A third tool improves kernel i-cache performance by moving "cold" basic blocks away from "hot" ones while the kernel is running (this is a little like tuning your car's engine while driving). This is joint work with Bart Miller, Alex Mirgorodskii, Nilu Motiwal, and Vic Zandy. slides
| Monday, | 9 October
| (No seminar)
| Monday, | 16 October 4-5 pm Room 2310
| Random Testing with "Fuzz": A Decade of Random Testing | Barton P. Miller In 1990, we published the results of a study of the reliability of standard UNIX application/utility programs. This study showed that by using simple (almost simplistic) random testing techniques, we could crash or hang 25-33% of the these utility programs. In 1995, we repeated and significantly extended this study using the same basic techniques: subjecting programs to random input streams. This study also included X-Window applications and servers. A distressingly large number of UNIX applications still crashed with our tests. X-window applications were at least as unreliable as command-line applications. The commercial versions of UNIX fared slightly better than in 1990, but the biggest surprise was that Linux and GNU applications were significantly more reliable than the commercial versions. This year (2000), we took another stab at random testing, this time testing applications running on Windows/NT . Given valid random mouse and keyboard input streams, we could crash or hang 45% of these applications. In this talk, I will discuss our simple testing procedure as it applies to UNIX and Windows. I will then present the UNIX and NT test results. These results include, in many cases, identification of the bugs and the coding practices that caused the bugs. In several cases, these bugs introduce issues relating to system security. The talk will conclude with some philosophical musings on the current state of commercial software.
software | papers Monday, | 6 November 4-5 pm Room 2310
| Matchmaking Failure Analysis
Nicholas Coleman
| The classified advertisement (ClassAd) language has been used to great success as a matchmaking framework for resource management in Condor. Providers and customers of services submit ClassAds containing their characteristics, constraints, and preferences to the Matchmaker. Ideally each customer should be matched with a provider such that both principals meet one anothers requirements. However, whether due to an incorrectly constructed ClassAd or overly strict constraints (from the provider or customer), some customers fail to find a match. At present there is no mechanism for providing matchmaking failure diagnostics in Condor. In this talk I will present several algorithms for matchmaking failure analysis which identify problematic aspects in a customer's ClassAd. I will discuss two separate (though not exclusive) cases of matchmaking failure:
Slides
| Monday, | 13 November 4-5 pm Room 2310
| Design of a Home Network | Rajesh Rajamani Networking technologies are starting to invade the home. The typical home network falls into two main applications area - Computer Interconnection and Control. The biggest market right now for home networking involves control - heating, ventilating, and air-conditioning systems. For mass market appeal, these networks will have to be inexpensive, easy to install, and the software easy to configure and operate. There is, of course, more to a network than just the physical and data link layers. Intelligent decisions have to be taken on the type of Operating system and communication protocols. I'll discuss some of the options that we had, the decisions we took, and the trade-offs we had to make, in designing and implementing a home-network at Synopsys's R&D facility in India.
| Monday, | 20 November 4-5 pm Room 2310
| Grid Security Infrastructure | Steve Tuecke, Argonne National Labs Grid applications pose numerous security challenges. The Globus Project (www.globus.org) has developed and deployed the Grid Security Infrastructure (GSI) over the past several years to meet these needs. This talk will give an overview of Grid applications, their security requirements, and how these requirements are being met by GSI. We will delve into the security protocols and APIs being used, application uses of GSI, implementation and deployment experience, and the future directions for GSI within the Globus Project and the Grid Forum.
| Monday, | 27 November 4-5 pm Room 2310
| Public Key Infrastructure Lab @ Wisconsin | Todd Tannenbaum
| |