Agenda | Biographies


Microsoft Research Faculty Summit 2004

Agenda for August 2 | August 3


Monday, August 2, 2004

Conference Introduction
Sailesh Chutani, Director of University Relations, Microsoft Research

Richard F. Rashid, Sr VP, Microsoft Research

William H. Gates, Chairman & Chief Software Architect, Microsoft Corporation

Innovation in SQL Server 2005 “Yukon”
Goetze Graefe, Microsoft

This presentation surveys a few specific areas of innovation in SQL Server 2005, from transaction processing to data mining, including the research that enabled the innovations and the benefits that our customer can derive from them.

Engineering Strategy at Microsoft
Jon Devaan, Senior Vice President, Microsoft

Implementing the CLR for Smart Personal Objects
Donald Thompson, Microsoft

Learn how Microsoft implemented an ECMA-compliant CLR runtime for ultra-small form factors, such as the SPOT wrist watch. The SPOT runtime is not a traditional OS, having been designed from the ground up with the goals of energy efficiency, constrained execution environment, over-the-air upgradeability, managed code development and tools (using Visual Studio .NET), and high reliability. This technology is currently being used in several internal projects and by Research, and we are exploring how to make it available for third-party development.

UI Innovations in Windows Longhorn
Pablo Fernicola, Microsoft

A look at the near-future Windows development platform, which will enable new experiences combining traditional elements from applications, documents, and media into common experiences. This talk will also focus on advances we will see in the next few years in relation to graphics and display capabilities and how they will also surface through the development platform.

Women of Microsoft Research
Sarah Revi Sterling, Microsoft

Tom Healy and Mythreyee Ganapathy
Alumni—IIT Bombay: Dr. Krithi Ramamritham, Dr. Uday Desai; MS Sponsor: Mythreyee Ganapathy
Alumni—NIT Warangal: Dr. DVLN Somayajulu; MS Sponsor: Tom Healy

Embedded RFP
Stewart Tansley, Microsoft
Ivan Joseph

This BOF is to gather together those professors currently working on Embedded RFP projects funded by Microsoft and discuss the final months of the project. There are 19 such professors currently registered for Faculty Summit, so we should have a good quorum. Specific items to discuss will include triaging any remaining technical or project issues, status check to confirm mutual understanding of how success is defined for the projects, discussion on the matter of fostering more of a community around the projects, and final workshop agenda and planning�how to present 78 successful projects at one event!

Tom Healy and Mythreyee Ganapathy
Alumni�IIT Delhi: Dr. Bijendra Jain, Dr. M. Balakrishnan; MS Sponsor: John Nordlinger/P. Anandan

Programmability Features in SQL Server 2005
Jose Blakely, Microsoft

The SQL Server 2005 product introduces many features designed to improve the productivity of database application developers. This presentation focuses on three main programmability areas: Integration of the .NET Common Language Runtime (CLR), native XML data type, and Service Broker. The integration of the .NET CLR inside the SQL Server DBMS enables database programmers to write business logic in the form of functions, stored procedures, triggers, data types, and aggregates using modern programming languages such as C#, Visual Basic, C++, COBOL, and J++. I will describe the key architectural aspects of this integration to provide a safe, scalable, secure, and efficient environment to run user code inside the database. I will show initial performance results demonstrating the efficiency of user-defined types and functions relative to equivalent native DBMS features. SQL Server addresses the semi-structured data management needs by implementing a native XML data type. I will describe key aspects of the XML index and integration with the relational query process. SQL Server addresses the needs of service-oriented application architectures by implementing Web-service access to SQL via HTTP, a service broker to handle asynchronous dialogs and messages, and query notifications.

A Web Service Architecture to Bring Labs Online
J. del Alamo
V. Harward
S. Lerman
K. Amaratunga

iCampus, the MIT/Microsoft educational alliance, has sponsored a campus-wide initiative at MIT to investigate the technology and pedagogy of online laboratories—real laboratory experiments that can be accessed through the Internet. This session will demonstrate a selection of lab experiments in several engineering disciplines that have been developed in this project. We will also describe a Web service architecture developed by the iLab team that has been designed to expedite the deployment and facilitate the management of online labs. The iLab approach, designed for cross-campus sharing of labs, splits the task of running experiments from student account management. This split frees lab providers from administrative chores and leaves faculty free to adapt the lab to their own pedagogy. This approach also leads to the identification of fully generic software modules that can be easily recycled across many types of iLabs. These generic modules will be released under an open source license so that others outside of MIT can more readily experiment with experimentation in their curricula. The presentation will also address the issues of sharing labs globally.

The Future of Web Services
John Shewchuck, Microsoft

This talk with covers Microsoft�s approach to Service Oriented Architecture (SOA), Web Services, and Indigo. We will provide an overview of some of the key research challenges Microsoft faces building our next-generation distributed platform.

Windows Embedded
Mike Hall, Microsoft

Microsoft has two embedded operating systems, Windows CE 5.0 and Windows XP Embedded SP1 (soon to be SP2). In this session we will take a look at the technologies supported by each of these operating systems; the tools that can be used to build, configure, download, and debug an embedded device; and how source and customization fits in with these operating systems. We will also take a brief look at the roadmap for Microsoft�s embedded operating systems and development tools.

Data Management Roadmap
Pedro Celis, Microsoft
Anil Nori, Microsoft

Managing data has become part of everyone�s daily life, both for enterprise users and for consumers. Maximizing the utility of data is key to realizing return on investment for enterprises, productivity for developers/knowledge users, and rich information experience for consumers. This session will present the Microsoft roadmap for end-to-end management of data. �Yukon� release of SQL Server and Longhorn/WinFS will deliver a storage platform for organizing, searching, sharing, and analyzing all types of data (for example, file, SQL, and XML), and for developing semantically rich application logic.

Piccolo .NET
Ben Bederson, University of Maryland

.NET application developers rely on Windows Forms for building GUIs, but what do we use to develop custom widgets, interactive graphics, and zoomable user interfaces? This talk introduces Piccolo.NET, an open library for building applications with rich 2D structured graphics. We will show applications demonstrating its utility and discuss its architecture and development paradigm—including a performance analysis and design implications. Piccolo.NET is also available in Java and for Pocket PC, and we will talk about the trade-offs of the different platforms and differences in approaches each requires.

Phoenix Academic Program
John Lefor, Microsoft

This talk will contain an overview of the goals, design, and status of the the Phoenix Compiler and Tools Framework. It will also provide updated information on how you can participate in the Phoenix Academic Program, which provides early access to Phoenix for use in your research or curricular activities.

Finding Similar Things Quickly and Compactly in Large Collections
Mark Manasse, Microsoft

It is often of interest, given a large collection of things, to quickly determine if many of the things are effectively the same. Consider, for example, the problem of identifying music from a Napster-like service: it may be of interest to RIAA to ascertain which of the songs are copies of songs in their catalog. It might be of use to a search engine for the Web to identify near-duplicate pages, to reduce the clutter on a results page, or to reduce the number of pages in the full-text index. It might be of interest to an intelligence agency when scanning a crowd at a public event, when trying to identify which members of the crowd are likely to be suspected malefactors. Note that in all of these, we are interested not in bit-for-bit identity (which is easy to determine using hashing), but in effective identity, despite variations in the source due to alternative ripping or compression, minor editing, or differences due to aging or imprecise measurements. This leaves us with the problem of finding the highly similar things in a large collection, or of identifying the most similar things in a large collection to a given thing. In general, we assume that there is a good algorithm for determining whether two given things are effectively the same. We assume that this algorithm works by feature extraction on each of the two things, and then by comparing features. We are interested in probabilistic algorithms that can select a subset of the features, compactly represent the selected features, and efficiently find candidates to consider for detailed examination.

Wireless Sensor Networks: Seamless Computing Across the Physical and PC Worlds
Feng Zhao, Microsoft

The proliferation of networked embedded devices, such as wireless sensors, ushers in an entirely new class of computing platforms. We need new ways to organize and program them. Unlike existing platforms, systems such as sensor networks are decentralized, embedded in physical world, and interact with people. In addition to computing, energy and bandwidth resources are constrained and must be negotiated. Uncertainty, both in systems and about the environment, is a given. Many tasks require collaboration among devices, and the entire network may have to be regarded as a processor.

The networked embedded computing group at Microsoft Research is developing new architectures, models, and tools for organizing and programming these systems, as well as innovative applications in areas such as security, transportation, and healthcare. Our goal is to build systems that are easy to use, manage, and program, robust to failures, and secure. We argue that the traditional node-centric programming of embedded devices is inadequate and unable to scale up. We need new service architectures, inter-operation protocols, programming models that are resource-aware and resource-efficient across heterogeneous devices that can range from extremely limited sensor motes to more powerful servers. In this talk, we will describe our ongoing work in developing a light-weight, multi-tier architecture that seamlessly mediates information flows between wireless sensors, micro-servers, and PCs, and discuss challenges in resource discovery, sensor tasking, load balancing, and uncertainty management.

Software Model Checking with SLAM
Sriram Rajamani, Microsoft

Model checking is a technique to find bugs in systems by systematically exploring their state spaces. Due to state space explosion, model checkers typically operate on a manually constructed finite-state abstraction of the system. The goal of the SLAM project is to model check software written in common programming languages directly. SLAM has been successfully used to find and fix errors in Windows device drivers written in the C language. SLAM operates by automatically constructing and refining model-checkable abstractions (called Boolean Programs) of a C program. I will describe how we have combined and extended results from the areas of program analysis, model checking and theorem proving to analyze critical properties of systems software written in the C language. SLAM is joint work with Tom Ball, Byron Cook, Vladimir Levin, and several interns and visitors to Microsoft Research.

Conference XP
Chris Moffatt, Microsoft
Michel Pahud, Microsoft

ConferenceXP is a shared-source, extensible research platform for designing and implementing collaborative learning applications. The goal of the project is to enable researchers to build less infrastructure and concentrate on the research, development, and deployment of applications that support and enhance collaborative learning in classroom and distance learning settings. By delivering support for high-quality, low-latency audio and video over broadband connections, sophisticated networking, and a collaboration framework and API, ConferenceXP does the heavy lifting to support the development real-time collaborative and conferencing applications that are compelling, scalable, and interoperable. The presentation will provide an overview of ConferenceXP, with an emphasis on its suitability to support wireless, collaborative scenarios in classroom and large lecture settings. The presentation will include a number of demos of innovative applications built on ConferenceXP.

The Verifying Compiler Grand Challenge for Computing
Tony Hoare, Microsoft

A Verifying Compiler is a programming tool that checks the correctness of every program that it compiles. Correctness includes generic properties such as security, crash-proofing, and deadlock freedom, as well as any application-specific functional properties (safety, serviceability) that the programmer may care to specify. Such a compiler would exploit, combine, and advance the state of the art in programming theory, program analysis, optimization, theorem proving, model checking, constraint satisfaction, design patterns, etc. A generic open-interface compiler like Phoenix could be a basic tool of this integrative research.

The development of a Verifying Compiler must be tested, guided, and evaluated by application to real programs that people actually write and use. The project I envisage would give equal emphasis to a scientific repository of challenge codes from many sources, ranging from critical embedded applications to commercial commodity products. The repository should include (in varying degrees) specifications, design documentation, test harnesses, test suites, existing proofs both manual and automatic, and the results of previous experiments conducted upon it. The highest credit must go to researchers who apply other people�s unconfirmed theories and other people�s untested tools to other people�s unreliable code, and contribute the results of a successful experiment to the repository.

An example of a similar Grand Challenge project has been the recently completed Human Genome project, which was a fifteen-year collaboration between laboratories in seven nations. I am currently enquiring if there is sufficient enthusiasm among researchers to mount a similar project in Computer Science. I am proposing an IFIP working Conference to be held next year in Zurich on Verified Software: theories, tools, and experiments.

Phoenix Tools Infrastructure: On-going Research from Three Early Adopters
Hoi Vo, Microsoft

Four speakers will present in this talk focusing on the uses of Phoenix in assisting research projects at University of AZ (Dr. Rajiv Gupta), Harvard University (Dr. Michael D. Smith), UC at San Diego (Dr. Brad Calder), and also within Microsoft Research (Hoi Vo, PPRC). In addition, we will provide information on the various applications of the Phoenix geared towards academic research, especially in the areas of Program Analysis, Optimization, and Instrumentation. Phoenix has been designed to run efficiently in a production environment while being flexible for research purposes.

Enabling Seamless HPC
Kyril Faenov, Microsoft
David Lifka, Cornell Theory Center

This session will present Microsoft�s product plans and roadmap for the High Performance Computing (HPC) market. High Performance Computing is one of the fastest growing server applications, driven by enterprises embracing server clustering and desktop cycle stealing to solve technical and business problems that only a few years ago required dedicated supercomputers. HPC features some of the most demanding and exciting application scenarios that drive innovation in distributed system development, large scale management, parallel computing, networking, and storage. This session will address how this exciting space is viewed and will be addressed by Microsoft—including product plans and roadmap.

Back to top


Tuesday, August 3, 2004

Daniel T. Ling, Corporate Vice President, Microsoft Research

A snapshot of Microsoft Research.


Birds Of A Feather Sessions

Shared Source of Common Language Infrastructure (SSCLI) in teaching Computer Science
Ranjan Sen
Jim Kavanaugh
Liz White
Witty Srinivasan

SSCLI is a free source-code distribution of an implementation of the Common Language Infrastructure along with C# compiler. It is a great resource for examination, testing, and experimenting within the context of teaching and learning many subjects of Computer Science. A brief presentation of SSCLI as a teaching resource will be given. A faculty from a Computer Science department will present how a debugger can be used to trace through the source of C# compiler in the SSCLI and help teach a course in Compiler Design. A faculty from another Computer Science department will present how Distributed Systems can be taught by exposing features of the Common Language Runtime through inspection and experimentation of the SSCLI (Rotor) code.

Lunch and Birds of a Feather Sessions

Grid Computing Using .Net and WSRF .NET
Marty Humphrey

The status of an implementation of the Web Services Resource Framework (WSRF) on .NET will be given. A summary of projects in grid computing and high-performance computing will be presented. This will follow a discussion on using WSRF in research and education

Tom Healy and Mythreyee Ganapathy
Alumni - BITS Pilani & India: Dr. Rahul Banerjee; MS Sponsor TBD
Alumni - IIT Kanpur: Dr. Sanjeev Aggarwal, Dr. Harish Karnick; MS Sponsor TBD

Stewart Tansley, Microsoft
Rich Davies

This BOF is to give you hands-on exposure to the new SPOT technology in a development kit form. The device�s program manager will be present to answer questions and get your reactions and feedback to this interesting new embedded system platform from Microsoft. It will be best if you attend the formal session presentation on the technology prior to attending this BOF as we will only provide a resume of the talk at the BOF itself and will want to dive straight into discussions!

John Lefor, Microsoft

This talk will contain an overview of the goals, design, and status of the the Phoenix Compiler and Tools Framework and will provide updated information on how you can participate in the Phoenix Academic Program, which provides early access to Phoenix for use in your research or curricular activities.

Interactive Photos and Video
Rick Szeliski, Microsoft

What will the future of interactive viewing of photos and video on the PC look like? Microsoft Research has been investigating and developing a number of new technologies in this area. These include photo stitching and interactive panoramic viewing, converting video clips into loops for backgrounds and screen savers, and creating real-world walkthroughs of beautiful environments (homes and gardens) using 360 video. Our most recent project is to create 3-D video that can be viewed and manipulated in real time. This talk will overview these technologies, complete with lots of visually interesting demos. It will also discuss open research challenges related to computer vision and image-based rendering in these areas.

Using Statistical Analysis to Locate Spam Web Pages
Dennis Fetterly, Microsoft

Commercial Web sites are more dependant than ever on being placed prominently within the result pages returned by a search engine to be successful. �Spam� Web pages are Web pages that are created for the sole purpose of misleading search engines and misdirecting traffic to target sites. Certain classes of spam pages, in particular those that are machine-generated, diverge in some of their properties from the properties of Web pages in general. As a result, these pages can be identified through statistical analysis. Microsoft Research Researchers have examined a variety of such properties, including linkage structure, page content, and page evolution, and have found that outliers in the statistical distribution of these properties are predominantly caused by Eeb spam.

Wireless Research of the Microsoft .NET Center in Taiwan
Prof. ChingYao Huang, Microsoft

In a cellular communication system, one of the important characteristics is the mobility. The mobility implies that each user might have a time-varying channel. This time-varying channel of each user allows the system to dynamically assign resources only to the users who are in better RF conditions subjected to QoS constraints. Even this user diversity could potentially provide significant performance gains, the benefit can not be realized unless the channel estimation and in-time resource management are also in place. To optimize system performance, channel estimation is a critical mechanism to provide necessary RF information of each user. In conjunction with in-time RF resource management, the system can then be optimized with a proper scheduler algorithm. Schedulers, as parts of radio resource management, are responsible for allocating RF resources to optimize system performance subjected to users� fairness and QoS constraints. In a real operation, fairness and a guarantee minimum rate are critical to maximize the number of satisfied users. With different data applications, the required minimum rates might also be different. As a result, the challenge of the scheduler designs is to optimize the system performance while achieving the expected level of the fairness and the minimum rate for all users. In this talk, we will first discuss the fundamental theory of achieving a high data rate system. A practical scheduler algorithm will be discussed to meet different design objectives.

Andrew Chien, University of California, San Diego (UCSD)
Jim Gray, Microsoft Research
David Lifka, Cornell Theory Center

Computing is not only becoming the technology fabric for business, but is also making major changes to how science is being conducted. This panel will highlight some of those changes and how modern computing technologies (databases, Web services, visualizations, etc) can still be applied. Some of the questions to be explored: How to deal with the data explosion? What other technologies can be applied? How can computer science researchers help this? With the real question being how can computer scientists help scientists focus on their science, rather than the computer science.

Research Directions in Social Computing
Shelly Farnham, Microsoft

In the Social Computing Group (SCG) at Microsoft Research, our mission is to research and develop software that contributes to compelling and effective social interactions. In this talk I will discuss recent trends in social software, and demo several recent SCG research projects designed to improve mainstream social practices that go both offline and online.

XML Web Services for Invisible Computing
Johannes Helander, Microsoft Research

Johannes Helander shows how XML Web services can solve the challenge of integrating invisible, embedded computers into larger systems to realize the ubiquitous computing vision. Web services were conceived to provide the interoperability required in e-business. An efficient implementation with a small memory footprint extends the paradigm to embedded systems and brings the benefits of a single messaging infrastructure to almost every class of computer and consumer device. The Microsoft Invisible Computing platform implements Web services on bare metal and is now available to academia. It provides a number of reusable and portable real-time system components. Recent work targets secure discovery, efficient XML processing, real-time C# execution, and distributed data-flow programming. Steve Liu from Texas A&M is using the Microsoft Invisible Computing platform to teach a course to 300 undergraduates and to perform his research in software-hardware co-design. He will present his pioneering experience with the new platform.

High-Quality Math Editing and Display
Murray Sargent, Microsoft

Infrastructures inside and outside of Microsoft enable us to enhance our programs to edit and display technical documents more easily and elegantly than ever before. A facility is described featuring display quality similar to that of the industry-standard TeX, a math input method similar to TeX�s but simpler and easier to use, and the ability to interchange formulae with popular mathematics programs such as Mathematica and MathCad.

Large Displays
Patrick Baudisch, Microsoft

The affordance of large personal screens is different from the affordance of regular sized screens. By extending further into the user�s visual periphery, large screens address the peripheral vision of the user and result in a natural distinction of screen space into focus and context regions. In this talk, we present a selection of interaction and visualization techniques that help users make better use of the additional screen resources offered by large screens and deal with some of the challenges brought up by these screens. Focus plus context screens turn regular screens into large screens by adding low-resolution peripheral screen space. Drag-and-pop is an interaction technique that provides users with access to peripheral screen content. Finally, we will investigate in how far lessons learned from large displays transfer back to regular and small screen devices. We will use the example of Halo, a visualization technique that supports spatial cognition by adding a virtual periphery to small-screen devices.

Pervasive Computing Research of Microsoft—HKUST Joint Lab
Prof. Qiong Luo, Hong Kong University of Science and Technology

Research on pervasive computing involves multiple areas of Computer Science, for example, graphics and vision, networking, planning, and systems. The virtual Microsoft-HKUST Joint Lab brings together Microsoft researchers and HKUST faculty members with a common goal of developing techniques that enable a wide variety of pervasive computing applications. In this talk, I will give an overview of our current research activities on pervasive computing. Specifically, I will introduce our ongoing work on interactive imaging tools, motion analysis of concentric mosaics, video multicast, application-level multicast, goal recognition in wireless local area networks, and query processing in pervasive computing and sensor networks.

The Next Media Group: Creating Novel Experiences for Interacting with Media
Steven Drucker, Microsoft

The NextMedia Group works on creating novel experiences for interacting with media. As opposed to focusing in on any one algorithm or technique, our group tries to envision new and compelling scenarios that integrate current research. Typically this process involves the creation of visionary papers, videos, demonstrations, and finally prototypes with which we can test the ideas and experiences.

In order to do this effectively, we need better ways of rapidly developing prototypes and iterating on the designs. This talk will overview some of the work done in the group over the last few years, ranging from work on prototyping tools, to many examples of the prototypes that we�ve explored. These prototypes range from systems for browsing a large database of videos, to other systems for annotating our personal photos, to ways of creating spectator experiences for video games.

Promises and Challenges of Community Wireless Networks
Victor Bahl, Microsoft

A recent study found that Internet use in communities� increased social interaction between neighbors, built up a community support structure, and improved the general well being of community residents. Community networks facilitate communications and information sharing locally. They are disruptive to the current residential broadband Internet access paradigm, which relies on cable and DSL being deployed in individual homes. They allow free flow of information without any moderation or selective rate control. Compared to existing systems that are centrally managed by ISPs, community networking is organic, where every participant contributes network resources and cooperates to form a self-organizing, self-managing mesh network. In this talk we discuss why this is an important new networking paradigm, parse the problem space, and discuss some important core technologies that Microsoft Research is building.

Trustworthy Computing: The Technical and Social Challenges Presented by the IT Revolution
Scott Charney, Microsoft

As we move into the information age, we find ourselves becoming increasingly dependent on computers, networks, and the information they contain. Yet each week we hear of new attacks on the confidentiality, integrity, and availability of computer systems, from the theft of personally identifiable information to distributed denial of service attacks. While such attacks may pose a serious threat to our public safety, national security, economic prosperity, and privacy, preventing and responding to such attacks raises other complicated social and public policy issues. Scott Charney will detail the history and current state of cyber threats, talk about Microsoft�s security initiatives, and identify the challenges governments, industry, and the public face as they attempt to prevent and respond to computer abuse.

Closing Keynote
Sailesh Chutani, Microsoft
Kevin Schofield, Microsoft

Community Discussion

Taking University Relations into the Future

Back to top


 > Events > Microsoft Research Faculty Summit 2004