1: Focused Effort Title: Ubiquitous Access to HPC Resources 2: Organization: Florida State University 3: Thematic Area: SciViz and DoD User Productivity 4: PI Name: Prof. Geoffrey Fox 5: E-Mail: fox@csit.fsu.edu 6: Telephone: (850) 644-4587 7: Fax:(850) 644-0098 8: Statement of Work: For the current year we are working on using Palm Top Devices as Discussed Below to Control and Monitor HPC Jobs -------------------------------------------------------------------------------- Given the work going on in "computing portals" world (activities supported by DoD, DOE, NSF, industry, etc.), it makes sense to envision the connectivity to HPC resources as services offered by a portal, and the handheld simply as a device with a particular set of capabilities that can link with the portal server. In this way the services don't have to be implemented separately for every different display device that comes along, and the portal can act as a flexible gateway to accommodate the different capabilities of various display devices. This project will demonstrate connectivity between HPC job and status information and handheld devices. In keeping with the portal/gateway concept described above, traditional web browser based access to the same information will be provided, but it will not be the primary focus of this project. Information available will include job initiation, progress, and termination (such as can be obtained from the batch queuing system). We will also investigate the possibility of including user-provided data (for example periodically providing the trailing few lines of the job's output file). Two-way interaction (such as user job cancellation and simple steering) will also be considered in the design, but will not be implemented unless we find a way to do this in general, taking into account security considerations. Six Month Extension ------------------- Palm-top and wireless technology is advancing so fast that it is impossible to predict what will be possible and "best to do" in this time period. I believe a critical development will be the deployment of Palm-compatible local area wireless technology such as Bluetooth. With this deployed, one will have companies like A T &T and Sprint offering ~14.4 kbaud linkage "anywhere" and then local environments like the MSU ERC or ITL lab would offer around one megabit Palm top connectivity; this hybrid performance is actually what one has with laptops today. This technology will be available generally in about a year but even if not useable during this performance period, we can use conventional 802.11 wireless in "high end" palm tops for demonstration projects. We believe that it is very important to understand how to use such devices in the the process of scientific discovery based on analysing computer simulations; interpreting the term broadly, this is the area of "Scientific Visualization on palmtops". I believe this a very important area and key to really revolutionary uses of Palm devices. As described above, we can expect that we will have good performance local area networking and one class of possibilities, we have identified (with CSIT visualization group Banks and Erlebacher) is using palm devices as intelligent controllers for large displays. This involves control of large screens and display of summary information (such as 2D slices) on the palmtop. We also see interesting generalizations of ERDC Kiosk -- for example we are installing many cheap SunRay display devices (or equivalent) on the walls of our research floor and will build software to allow palm control of such "pervasive" displays. This we think could change the way we do computing (and looks towards time whole walls can be tiled with displays). As a final example, we are modifying the public domain VNC software to share thumbnail versions of conventional displays on palm tops. This would allow quick inspection for gross anomalies and perhaps nifty icons to label notes taken on palmtop. We view these as exemplar projects to be refined as a basis of our early experiments which we have already started. The ERDC sponsored "Hand Held Supercomputing" workshop to be held february 2001 will be very helpful in making decisions. We have identified some really excellent academic and industry contributors 9: Deliverables: Use of Palm Top Devices to Control (Large Screen) Displays Collaborative thumbnails shared on Palm Tops Written progress reports in June 2001 Presentations for PET Final Review in September 2001 Final Technical Report in September 2001 10: Required Resources: $48,291 Item Base Fringe Overhead Total Geoffrey Fox 7,500 1,380 4,129 13,009 2 Graduate Students 16,000 96 7,485 23,581 (.5 year) Tuition 2,306 2,306 Travel 3,000 1,395 4,395 Hardware 5,000 5,000 TOTAL 48,291