Abstract

The DART project, a collaboration between Monash University, the University of Queensland and James Cook University, has been exploring the application of eResearch tools and middlewares in support of scientific workflows.

As a specific example of an ‘end-to-end’ eResearch workflow we have developed an implemented a remote access regime for accessing both small molecule and protein diffractometers, coupling this system with a federated data management and portal interface system. Our goal has been to provide an end-to-end solution for protein and small molecule crystallography. This system, called JAINIS, is based on the Common Instrument Middleware Architecture (CIMA) architecture and uses a seperate computer to provide proxy access to the scientific instrument (in this case an X-Ray Diffractometer) termed the instrument representative (IR). A separate (and possibly remote) data manager then receives data from the IR, as well as video streams that monitor the actual laboratory environment. We have enhanced the CIMA model so that data (including CCD image frames) is stored in a secure, federated Data Grid storage infrastructure (SRB or storage resource broker). A workflow system (Kepler) is being used to orchestrate the data flows within the data manager, and to automate data storage, extraction and generation of instrument and experimental metadata and provide the triggers to alter instrument operators of instrument status.

Access to JAINIS is via a web portal, and provides a customisable environment that gives users access to live data from running experiments (including latest CCD and microscope images), an ambient view of the laboratory environment as well as access to historic experiments and their data. Comprehensive data searching as well as file uploads and downloads are also supported. Data stores from different sites can be federated to provide a single virtual data repository, and a security and right management capability are now being implemented using shibboleth based architecture.

About the speaker

Ian Atkinson has a background in computational chemistry that led him to the world of Supercomputing. More recently he has been engaged in the APAC Grid program providing portal and grid computing services to the national research community. He has a long-standing interest in eResearch methods and applications including shared visualisation via AccessGrid and in particular grid mediated access to data sources. He is a CI on the DEST DART eResearch project focussing on data collection from instruments and sensors.

He holds joint appointments as the Manager of High Performance Computing within Information Technology and Resources at JCU and also a facility member of the School of IT at JCU. He is also the Deputy-CEO of QPSF.