NSF IDM PI Workshop Research From An Industry Perspective Raghu Ramakrishnan UW-Madison (and QUIQ … IBM/Microsoft/Virage)
Research Spectrum Small Big Very small Contract Grant INDUSTRIAL ACADEMIC
Research Climate:Academia Where do I find interesting problems? –Make them up –Solve someone else’s (Collaboration!) Joint projects with industry Where do I find money to solve them? –If you find out, let me know How do I find students to solve it for me? –By becoming famous, writing a lot of papers (and having money)
Research Climate: Big Company Where do I find interesting problems? –The product groups in the company –Customers Where do I find money to solve them? –My manager –Product groups –Customers How do I get to hire more people? –By making those with money happy
Comments Shift in financial conditions is forcing shorter horizons for research, and increasing demand for near-term returns –Product groups vs. research groups dynamics Continued interest in working with academia, but less money to support academic partners
Current Areas of Interest Manageability, diagnostics, tuning On-demand computing (“grids”, clusters) XML Text Analytics Data cleaning, integration Security, privacy Applications (bio, streams)
Research Climate: Medium Company Where do I find interesting problems? –You don’t find them. You solve problems you have to, and focus on building a product. Where do I find money to solve them? –By selling your product How do I get to hire more people? –By building and selling a product
Comments Hard to find ways to collaborate with this category of company –But don’t let that stop you from trying—I’m just sharing one perspective, and your experience may be different
Research Climate:Very Small Company Where do I find interesting problems? –Refining your company’s core IP –Reading a BAA Where do I find money to solve them? –Federal contracts, SBIR grants How do I get to hire more people? –No, you try to pay the people you already have.
Comments Collaboration is typically by invitation— insiders drive the work, and call on people they know to fill in gaps –Send Chrismas cards to all your friends who just started a small company
NSF IDM Workshop Science of Design for Information Systems SDIS 2003 Phil Bernstein Alon Halevy Raghu Ramakrishnan
Workshop Goal Identify research challenges in design of information systems –Unique opportunities, motivating scenarios Identify disciplines and groups that must work together to solve these challenges –Across academia and industry Recommend NSF activities to support SDIS –What types of activities, grants?
People A: Design and Data Integration –Bernstein, Borgida, Florescu, Gasser, Halevy B: Design for Dynamic Data Environments –Chaudhuri, Franklin, Ramakrishnan, Tsotras, Widom, Yuan C: Design and Data Location –Ghandaharizadeh, Jagadish, Korth, Srivastava, Wolfson D: Designing around complex relationships, processes –Carey, Chrysanthis, Dayal, Su, Thatte E: Designing for Security and Privacy –Clifton, Gehrke, Rosenthal, Sripada, Suciu, Winslett At large –Maria Zemankova, Bhavani Thuraisingham
Group A A: Design and Data Integration –Bernstein, Borgida, Florescu, Gasser, Halevy Semantic integration of legacy enterprise systems and processes –Deploying complex off-the-shelf systems –Managing models and mappings –Emerging standards, services Novel settings –E.g., Internet-scale semantic integration; Bioinformatics, medical informatics; large-scale collaboration
Group B B: Design for Dynamic Data Environments –Chaudhuri, Franklin, Ramakrishnan, Tsotras, Widom, Yuan Rapidly changing data –Continuous data acquisition, processing –Information from ongoing analysis Provenance of data, processes, schedules Conflict-resolution; version management Data aging and archiving
Group C C: Design and Data Location –Ghandaharizadeh, Jagadish, Korth, Srivastava, Wolfson Middle-tier caching and replication Highly distributed data collections –Web, P2P Mobile environments and objects –Monitoring, tracking, and reactive systems –Personal computing
Group D D: Designing around complex relationships, processes, interactions –Carey, Chrysanthis, Dayal, Su, Thatte Extend the scope of design beyond the traditional focus on data, to encompass patterns of activity –Business rules, workflow – Application-level interactions, web services
Group E E: Designing for Security and Privacy –Clifton, Gehrke, Rosenthal, Sripada, Suciu, Winslett Fine-granularity access control, especially in dynamic systems User-level preference specifications Methodology of how to develop information systems that utilize emerging tools and verifiably achieve security and privacy goals –E.g., Secure, privacy-preserving analysis –E.g., Preventing and detecting intrusions