RETAIL

MOUNTAIN VIEW, Calif. -- Indicating the growing acceptance of the SGI Visual Area Networking concept, the Montreal Neurological Institute (MNI) is the 100th user of the year-old technology. The institute's deployment of Visual Area Networking is powered by a five- pipe SGI(R) Onyx(R) 3400 visualization server with an InfiniteReality(R) graphics subsystem and OpenGL Vizserver(TM) client/server software (purchased in 2002). This compute infrastructure enables collaborative work with extremely large data files over very long distances, which is becoming a key requirement for neuroimaging researchers. The SGI technology is a key component of several long-distance collaborative neuroimaging efforts the MNI is undertaking worldwide. These include: -- A planned project between the Julich Research Centre in Julich, Germany, and Massachusetts General Hospital in Boston -- A demonstration project between Los Angeles and Sendai, Japan, as part of the International Consortium for Brain Mapping -- A visualization network of institutions in the greater Montreal area and the province of Quebec Prior to installing the OpenGL Vizserver software, MNI participated in long-distance testing of the technology, led by SGI, in which visualizations of data sets were successfully transmitted to sites extending from Montreal to as far west as Calgary (1,900 miles) and as far east as Halifax (500 miles). OpenGL Vizserver allows MNI's remote partners to access Montreal-based data via commodity networking infrastructure. It compresses the requested image generated on the graphics subsystem and transports that image to the client, which then decompresses and displays the image. The nascent Quebec network is one of the foremost examples of how neurological researchers can share their work without having to move extremely large data sets, which has been a severely limiting factor in the computationally intensive brain-imaging field. Once the network is fully operational, data will continue to reside in databases at MNI, and researchers at both MNI and remote locations will be able to view and manipulate them via the visualization network. "We're starting to work with groups all over the world," said Alan Evans, Ph.D., director of the Montreal Consortium for Brain Imaging Research and professor of neurology at McGill University, home of MNI. "Increasingly, we're dealing with really big files. Our colleagues in Germany, world renowned for their work on cadaver brains, are producing data sets that range up to 800GB, huge things. Those can't be sampled by any conventional laptop or workstation." The SGI Onyx 3400 system and OpenGL Vizserver software at MNI will be integral components for the planned collaboration with researchers at Massachusetts General Hospital. They will scan the brains of drug addicts and send the resulting data to MNI's data storage facilities, from which researchers at any remote partner site can access and view it with minimal network latency using OpenGL Vizserver. Utilizing the ultrahigh-speed broadband CANARIE and RISQ networks, built by the Canadian federal government and the provincial government of Quebec respectively, researchers at MNI and remote sites worldwide will be able to simultaneously view and manipulate extremely large data sets using such client devices as SGI workstations, and Windows(R) and Linux(R) OS-based PCs. Up to five separate collaborative sessions can be conducted simultaneously on the Onyx(R) system at MNI. Michael Brown, SGI marketing team manager, advanced graphics, points out that the MNI deployment demonstrates that burgeoning power of Visual Area Networking across a broad array of data-intensive uses. "We are seeing a really broad adoption of Visual Area Networking in government, education and industry," Brown said. "The ability for end users to visualize large amounts of data from their offices is critical for supercomputing centers and distributed grid computing environments, where users can be certain that their data will not be local and that their desktop systems won't be able to analyze it. The MNI environment shows how a visualization-centric approach to computing can accelerate data access and collaboration within a global community." The installation of Visual Area Networking technology is the latest chapter in a long and successful collaboration between SGI and MNI. "It's a perfect marriage and has been for many years between our group and SGI, in terms of joint development," said MNI's Evans. "We got involved with SGI in the late 1980s and early 1990s because we needed 3D graphics, and SGI was the best in the business at the time. As our field evolved, we became interested in very high levels of computing, whether this was data sets in the thousands, really big data sets, or both. SGI at the time acquired Cray and started working in that domain, which suited us perfectly. Then, as we started moving into multicenter projects, Web-based communication between sites became a big priority, and SGI was there, waiting for us again with OpenGL Vizserver."