The XPort Project
 
 

An NGI Application Testbed Partnership BetweenIndiana University and Argonne National Laboratory: A Grid-based Collaboratory for Real-time Data Acquisition, Reduction and Visualization for Macromolecular X-Ray Crystallography Using the LBL Advanced Light Source

Randall Bramley, Principal Investigator
Donald F. (Rick) McMullen, Project Director
Indiana University

Ian Foster, Principal Investigator
Gregor von Laszewski, co-PI
Argonne National Laboratory

Project Description:
This project targets revolutionary improvements in telepresence for major scientific instrumentation systems.  Our goal is to exploit a combination of advanced networking technology, sophisticated middleware services, and remote instrumentation technologies to achieve interactive "better-than-being-there" capabilities for remote experiment planning, instrument operation, data reconstruction, and data analysis.  These capabilities will be deployed and demonstrated at major DOE facilities, including the Advanced Photon Source and the Advanced Light Source.  Building on work currently underway in the Globus group and the DOE 2000 Common Component Architecture Forum, this proposal addresses several issues related to NGI network-based instrumentation including high speed data collection, reduction, storage and visualization, and real-time instrument control for the acquisition of macromolecular x-ray crystallographic data from the MB-CAT beamline sector at the LBL Advanced Light Source.

Indiana University and the Globus group at Argonne National Lab are developing this Next Generation Internet (NGI) based shared instrumentation collaboratory for macromolecular crystallography.  This partnership also includes researchers in the DOE2000 Common Component Architecture group (CCA) and in the Molecular Biology Collaborative Access Team (MB-CAT) at the Berkeley Advanced Light Source (ALS).

The unique scientific instruments operated by DOE laboratories are extremely visible, expensive, and scientifically valuable; unfortunately, they are also frequently highly inconvenient to use. For example, in order to use the APS or ALS, scientists must wait months for beam time, then travel thousands of miles and spend many days on site in order to collect data.  Once data is collected, days or weeks may be spent on reconstruction.  Hence, time to information is months rather than the hours, and the duty cycle of the instruments themselves is poor.  Our work is centered on letting the scientists submit jobs remotely, view the session interactively, receive and analyze preliminary results to validate the sample being tested,  to store and manage the data produced, and to interactively show and discuss the results with collaborators.  This work will have major impact on U.S. science and industry, by improving both time to information and duty cycle for major DOE facilities. Because time to information is reduced from months to hours and interactive use of the remote facility is possible, once specialized resources can be integrated into the laboratory of the ordinary scientist.

Another area where telepresence and grid computing can have a significant impact on productivity and the quality of work done is during the analysis of data and model construction by extended collaborations.  Grid based virtual workspaces, such as the ANL Metro project, can bring collaborative or consultative teams together with relevant data efficiently in real time or asynchronously.  Providing analytical and modeling tools for crystallography in shared virtual workspaces is an expected outcome of the proposed work.

Several projects have taken important steps towards the virtualization of major instrumentation.  For example, the DOE2000 Materials and Microcharacterization Collaboratory (MMC) is exploring the use of collaboratory technologies to support remote use of microscopes and other devices, while the DOE X-ray source Grand Challenge project is investigating real-time supercomputer reconstruction and image analysis of X-ray source data.  However, these and other efforts must be characterized as exploratory: true telepresence and remote operation are far from being achieved.

Other instrument collaboratories based on emerging NGI infrastructure and Grid computing tools include:

• the National Center for Microscopy and Imaging Research Collaboratory for Microscopic Digital Anatomy involving remote use of ultra-high voltage electron microscopes,
• the Distributed Parallel Storage System (DPSS) used in support of high energy physics experiments ,
• the Wide Area Large Data Objects (WALDO) distributed digital library system used in support of cardioangiography and other real-time medical imaging applications.

• an NGI (ESNet/Abilene/vBNS) hosted testbed for macromolecular crystallography;
• a component-based application development environment for instrument control and experimental data collection. The Common Component Architecture Forum's specifications and current deliverables form the basis for development of the proposed distributed data collection, reduction and storage system;
• high-speed data caching and transfer middleware to support direct connection of the ALS instrumentation to grid mass storage and computing systems;
• distributed real-time data reduction components which can be run during data collection to determine quality of the data being collected and to provide feedback for instrument tuning;
• extensions to the current Component Architecture Toolkit (CAT) to enable quality of service (QoS) and real-time network performance analysis for a production application.  These are needed to optimize local caching of data and network bandwidth use during different phases of an experiment. and to guarantee data integrity in a tightly coupled data reduction and analysis pipeline;
• extensions to the CAT component toolkit to allow interactive performance monitoring and analysis of computation and communication in individual components;
• experiment-specific resource discovery, authentication, authorization, allocation, and co-scheduling shims based on the Globus toolkit for the integration of the ALS data acquisition hardware, network, processing, storage, and visualization elements.

The MB-CAT testbed and other distributed instrumentation applications depend on the allocation and co-scheduling of a geographically diverse set of hardware and software resources.  We use the Globus toolkit to provide appropriate middleware on which to compose component-based software in support of specific experiments and routine tasks associated with the MB-CAT's beamline sector.

Related Links

• The Indiana University Molecular Structure Center (IUMSC)
• The IUMSC WWW Distribution Site, which has a brief description of X-Ray crystallography and a server for structures determined by the IUMSC.
• The Indiana University Common Component Architecture Toolkit (CCAT), the primary software framework for the XPorts project.
• Project personnel and contact information
• DoE NGI project characteristics summary table
• PowerPoint presentation, October 1999 NGI PI meeting
• XPort architecture and components
• Initial XPort software configuration picture
• Local documentation
• SC2000: Poster, Virtual poster session, Talk slides