Information Technology //
MRN's computing infrastructure consists of approximately 250 computers and ~660 Terabytes of network attached storage that consolidates all research data to facilitate easy access and retrieval. Desktop workstations are fed by Gigabit Ethernet connections to a high availability data center that houses core services as well as multiprocessor compute servers.
Scientific end users are able to access the MRN resources using a single sign on environment that enables sharing of data across the Windows, Linux and Macintosh platforms. The data center houses several large shared memory systems (~20 cores & 128GB RAM on average). Additionally there are currently over 30 image analysis and neuroinformatics applications installed and administered uniformly across all workstations. Primary data analysis packages are MATLAB, SPM, GIFT/ICA, AFNI, FSL and FreeSurfer. FreeSurfer currently runs on one of MRN's 32 processor clusters.
The data center is backed by a 30 KW N+1 UPS Uninterruptible Power Supply (expandable to 80KW). Data center air conditioning is protected by a diesel generator for emergency power and all power and cooling systems are monitored by IT staff 24/7.
All desktop workstations communicate with MRN's data center and network attached storage via dual redundant fiber links. MRN is connected to the Internet via gigabit fiber to a 100 megabit per second link on the University of New Mexico's campus. A Virtual Private Network allows researchers and collaborators to access MRN computing resources remotely through an encrypted link.
Data acquisition computers across all modalities (1.5T Mobile, 3T, MEG, EEG) have standardized stimulus delivery systems i.e. pupilometry, audio, video and time coding. The mobile data acquisition system includes a group of centrally managed, workstations used for psychological assessment, digital video and document imaging in forensic populations. The network attached storage systems are a combination of high availability EMC and Sun filer storage pools totaling ~660TB of usable disk space. Research data is stored in a self documenting directory hierarchy (both raw and processed).