PS: The authors of this paper include: Walter Sadowski - Department of Energy Vincent Chan and Bill McHarg - General Atomics Thomas Casper and James Crotinger - Lawrence Livermore National Laboratory Martin Greenwald - Massachusetts Institute of Technology Plasma Fusion Center David Greenwood and Wayne Houlberg - Oak Ridge National Laboratory Dori Barnes, Steve Davis and Steve Jardin - Princeton Plasma Physics Laboratory John Cary - University of Colorado at Boulder _____________________________________________________________ ______ | ___ \___ Dr. Steven L. Davis, Head, Systems Development (609)243-3170 | ____/__ \___ Princeton Plasma Physics Laboratory FAX:(609)243-3086 |_| | ____/__ \ P.O. Box 451 |_| | ____/ Princeton, NJ - 08543 |_| | |___ |_____| sdavis@pppl.gov _________________________________________________________________________ National Magnetic Fusion Collaboratory DOE Office of Fusion Energy Science General Atomics Lawrence Livermore National Laboratory Massachusetts Institute of Technology Oak Ridge National Laboratory Princeton Plasma Physics Laboratory University of Colorado at Boulder INTRODUCTION This white paper presents plans for the creation of a National Magnetic Fusion Collaboratory. The collaboratory is a response to the increasing dependence on a few major experiments, a more critical reliance on computational modeling, the need for collaborative code development and the internationalization of the magnetic fusion effort. This proposal is based on the white papers "Requirements for a National Fusion Collaboratory" [1], which has been extended in scope to include international collaborations [2], and "Needs and Opportunities in Computational Fusion Science" [3]. OVERVIEW We propose to implement a National Magnetic Fusion Collaboratory (NMFC) designed to increase efficiency of research efforts in the US magnetic fusion program by means of facilities and personnel sharing. The NMFC will allow researchers around the country to participate remotely in experiments and applications code development using common tools and interfaces. Our proposal is based initially on a consortium involving six sites in the US with access to three tokamak experimental facilities: Alcator C-Mod, DIII-D, and NSTX. Several applications codes that are currently used nationally and internationally, as well as one or more new codes to be developed, will also be part of the NMFC. We will deploy the tools needed to establish a collaboratory. These tools will be tested on a heterogeneous network, with integrated control room and office environments available to participants. Audio/video conferencing technologies will be used for remote monitoring and meetings. Electronic notebooks, whiteboards and shared desktops will be used in applications such as experimental recording, code development and documentation. Experimentalists will exercise remote control of diagnostic instrumentation, heating and fueling systems and other experimental parameters up to and including full functional control of the tokamak. Appropriate authentication techniques will be implemented to ensure security of networked resources. PRIOR EXPERIENCE IN REMOTE COLLABORATION The fusion community has considerable experience in implementing and using remote collaboration tools. Recent collaborations in the fusion community have received widespread recognition. Remote control of Alcator C-Mod from LLNL [4] was highlighted in articles in Business Week (October 30, 1995, page 68), Science (April 7, 1995, Vol. 268, page 35), PC Week (August 7,1995), Control (June 1995, page 13) and Science and Technology Review (January/February, 1996). The REE project was named as a 1996 finalist in the "Next Generation" category for the 1996 National Information Infrastructure (NII) award. OBJECTIVES Our overall objective is to allow scientists at remote sites to participate as fully in our experiments and computational activities as if they were working at a common site. These objectives can be realized through a collaboratory that consists of five components: (i) Remote Experimental Operations The key technologies to support remote participation are those that enable secure, transparent, timely and platform independent access to: information on instrumentation and experimental status; raw, analyzed and modeling data; and those technologies that permit remote scientific discussions. These technologies should form a comprehensive environment suitable for both physical and numerical experimentation. We will test and implement various options for remote, distributed access to code, files and data. We will include fully authenticated read/write access to data and appropriate interfaces to allow remote control of experimental systems and diagnostics. High quality audio and video communications among the control rooms and researchers integrated into the existing desktop video displays or external displays are essential. Early emphasis will be placed on high-quality audio and the ability to control and monitor access. (ii) Multi-Institutional Code Development Projects We propose to make use of modern collaboration technology to enable multi-institutional teams of experts to participate in joint code development efforts. Each team, involving individuals from three or more sites, will make effective use of desktop video conferencing, a common code-version control system, an electronic notebook to document coding changes, common display output demonstrations and other available communication and collaboration technologies. (iii) Codes On-Line The NMFC will implement a common on-line code-sharing library to improve code and data access, code interconnection and code invocation. This on-line facility will include codes for both modeling and data analysis. Construction of this resource will be a phased multi-year effort. In the first phase we will adopt and implement I/O standards to facilitate both the transfer of experimental data into the analysis codes and the transfer of processed output data among the codes. In this phase we will also develop network-based tools for code and database invocation (local or remote) and visualization. In subsequent phases, we will address such issues as the efficient use of distributed compute cycles, reliable asynchronous intertask communications, multicasting data, remote display and downloading of results, distributed task queuing and session management. We also expect to address the ability to select and run related codes as a unit, or to dynamically link code modules in a scripting or software bus framework. (iv) Planning and Coordination Group interactions for planning and coordination will take place over the network, using internet-based desktop video and audio technology. For our multi-institutional applications, considerable work is still needed to improve desktop teleconferencing tools. We require improvements in video quality and hands-free, full duplex audio. Local facilities will be upgraded to provide these performance enhancements. Informal audio/video "chat corners" will be set up for impromptu scientific and technical conversations between collaborators. Open meetings, both large and small, will be broadcast over the network. We will explore ways to make remote participation in meetings more interactive and we require improved remote access to both manually and electronically prepared presentation materials. Network based audio/video recording and playback tools are needed to accommodate national and international time-zone differences. (v) Sociology of Extended Environments Understanding is needed of the relationships between people and their work environment, especially over large geographical regions and including extensive implementation of communications technologies. An on-going sociological study will help to define some of the directions or modifications to the technical environment and identify work and management practices to promote more productive collaborations. REFERENCES [1] D. Barnes (PPPL), T. Casper (LLNL), S. Davis (PPPL), M. Greenwald (MIT), D. Greenwood (ORNL), P. Henline (GA), J. Moller (LLNL), P. Roney (PPPL), "Requirements for a National Fusion Collaboratory," white paper, October 1996. (See: http://www.pppl.gov/nfc) [2] A. Katz (DOE) et al., "Requirements for Remote Participation in Fusion Research Using Telecommunications: A Report to the Fusion Power Coordinating Committee of the International Energy Agency by the Working Group on Remote Participation," currently in DRAFT form. (See: http://wwwofe.er.doe.gov/More_HTML/remote.pdf) [3] B. Braams (NYU), J. Cary (U. Col.), V. Chan (GA), R. Cohen (LLNL), W. Dorland (U. Texas), W. Houlberg (ORNL), S. Jardin (PPPL), D. Schnack (SAIC) "Needs and Opportunities in Computational Fusion Science," white paper, October 1996. (See: http://cips.colorado.edu/NeedsOppsCompFusion.html) [4] S. Horne, M. Greenwald, T. Fredian, et. al., "Remote Control of Alcator C-Mod from LLNL," to be published in Journal of Fusion Technology (1997). (See: http://cmod2.pfc.mit.edu/~g/remote.html)