CRCU Resources
Space
The CRCU is housed in approximately 7,500 square feet of space on the 5th floor of the 3535 Market Street building. In addition, the IST support team is housed in approximately 650 square feet of office space on the first floor and 400 square feet of machine room space on the fifth floor of Blockley Hall. This space is allocated to the CRCU for staff offices, machine room space and operational/project function area in support of major new initiatives in large-scale, multicenter clinical trials, clinical and patient-oriented research projects, and multi-institutional health services research projects.The Market Street space was renovated for the CRCU and designed specifically to encourage collaborative project team activities. The space includes offices for the directors, individual offices for technical group leaders, a receptionist area, and open seating areas designed to accommodate multiple technical personnel. This 5th floor area also provides one conference room equipped with appropriate A/V technologies supporting computerized demonstration and presentation techniques. Large central-file areas have also been strategically located and designed to accommodate the movement and storage of case report forms, manuals of operations, and clinical data management materials and documents. A special document handling area for processing large volume copying, scanning, and document mailing and postage activities normally associated with large-scale, multicenter clinical trial projects, has also been provided.
In Blockley Hall, two specially designed machine rooms are also part of the CRCU computing environment. These rooms are equipped with UPS power conditioning, special air cooling and handling, and discrete computer network capabilities, all designed to ensure maximum availability and protect the computer technologies used by the CRCU.
Computing Resources
Overview:
The CRCU computing environment consists of commercially available cutting-edge technologies. All technologies used by the CRCU are production quality and fully supported by the respective vendors. This ensures that the most current vendor-supported technologies are used for the project. The technologies fall into one of several areas:
- The communications network that supports all the computing and communications for projects developed and supported by the CRCU.
- The hardware configurations on which all applications and software are built and supported.
- The application and system software used to develop, maintain, and analyze the data and processes for the project.
Communications Network
The communications network consists of a switched TCP/IP Ethernet environment composed of several intra-networks. The server intra-network connects multiple computer servers to the rest of the CRCU network, providing a minimum of 100 megabits (Mb)/sec of network connectivity to each server. There is also a client intra-network consisting of a 10Mb/sec TCP/IP switched Ethernet network. Both the server and the client intra-networks are connected to the University of Pennsylvania's Ethernet network through a dedicated 10 Mb/sec-network connection.Hardware
- Core CRCU Servers: The core CRCU servers consist of 15 Unix servers supporting several user and database services for the various entities and research projects supported by the CRCU. Outlined below, these servers have a diverse set of functions such as mail, network and user file and directory services, printing support, and security.
- Statistical and Data Management Servers: The servers used for the statistical and data management support on projects consist of SUN Microsystems E4000 Enterprise servers, SUN 280R compute servers and SUN Microsystems E250 servers. The servers are also configured with Redundant Array of Inexpensive Disk (RAID) subsystems to support the file and directory needs of the project work and staff requirements. The total disk space currently available to these servers is over 200 gigabytes (GB) of high-speed disk configured using RAID0+1 and RAID 5 specifications. New hardware is being readied to replace the core computing resources, and will include a Sun Grid Engine arrangement in which eight Sun v440 servers running Solaris and three Sun v20 servers running RedHat Linux comprise a single High Performance Computing environment.
- Database Servers: The servers employed for the ORACLE databases used for projects are SUN Microsystems E4000 Enterprise servers, used for the development databases, and a SUN Microsystems E250 for production databases. The total disk space available for these servers is over 250 GB of high-speed disk configured using a RAID5 specification. Again, new hardware is being configured to replace all of the above database servers. A Sun v440 server with four processors and 16 GB of memory will be the new production database engine and available disk space will be 1 terabyte.
- Servers for Workgroup and Network Services: In addition to the primary servers for statistics and database processing, there various ancillary servers that provide the workgroup services as well as network and security services for the CRCU. There are multiple SUN E250, SUN Ultra-1, and SUN Ultra-5 class servers configured to accomplish these tasks. These servers provide the network services of NIS+ and NFS, automatically mounting home and project directories while maintaining the authentication and encryption mechanisms required for the CRCU network and data. These servers also provide primary and secondary functions for electronic mail, directory services (LDAP), calendar and scheduling, nightly backup services, and other network protocols. The network is also provided a layer of protection from outside traffic through a "firewalled" gateway providing dynamic packet filtering.
- Storage Area Network (SAN): Another technological advancement for this fiscal year is the addition of a Storage Area Network. With a starting capacity of 4 terabytes, the SAN delivers a scalable architecture for providing disk to the High Performance Computing grid, the Oracle database server environment, and all other project and infrastructure demands. RAID 5 disk groups with "hot spares," dynamic multipathing fibre connections to servers, automated vendor notification of any hardware failures, monitoring software, and administrative tools make managing rapidly increasing storage a centralized, highly available resource.
- Client Platforms (400+ boxes): The CRCU is a heterogeneous computing environment. The core server facilities support Unix- and Intel-based processes. The client workstations used throughout the CCEB must be able to support the protocols required by the computing environment and the project. All data are centrally stored on server-attached disks (2 TB of disk) rather than individual client disks.
- Backup Operations: Since data are stored centrally, backup operations are also centralized. A robotic tape library containing dual tape drives and 32 AIT-3 tape cartridges provides nightly protection against data loss, and the ability to create duplicate tapes of monthly archives for storage off-site as part of a disaster recovery plan.
- Printing: Printing is accomplished using various laser printers. Departmental printers include several HP LaserJet 4100TN printers and an HP Color LaserJet 5M and 8500. Numerous work group printers (HP 6MP or higher grade) are also available for project use. All of these printers are networked and located throughout the CRCU facilities for project printing needs.
- Operating Systems (OS): The operating systems for the SUN Microsystems servers and any client Sun Microsystems workstations are Solaris 8 and Solaris 9. The OS for the Intel-based clients are either Microsoft Windows 2000 or XP, and servers are Microsoft Windows 2000 and 2003.
- Database Software: The database management and development software standardized for use within the CRCU is ORACLE. Along with the ORACLE database management software, there are several developmental tool kits from ORACLE to provide World Wide Web-based development and data repository software development. These tool kits allow development of modular and reusable code segments to enhance and expedite the applications developed within the CRCU. We are looking to use Oracle Clinical software as our standard and are currently developing an environment for production use in the next fiscal year.
- Statistical Software: Sundry statistical software packages are available to provide statistical analysis for project data. Software supported by the CRCU includes SAS, SPLUS, SUDAAN, GAUSS, STATA, STATXACT and EGRET.
- Productivity Software: Key productivity tools available to all members associated with a project team consist of e-mail client software, Netscape Calendar and Scheduling software, word processing software in the form of Microsoft's Office Suite, Adobe Acrobat and Adobe FrameMaker, and the TeX/LaTeX document processing language. These tools support the broad spectrum of tasks on a project. Diagrams and flowcharts are produced within the Visio Graphics software package.
