Computing reality • We tend to be dealt the computing environment in which we must operate. • Few of us have enough influence to steer the direction of central computing on our campus. • Thus, we try to match our computing needs with the resources provided locally.
Match to level of service Develop a computing strategy that identifies • the hardware, • the software, and • the network connectivity needed to support the level of data service you are providing now and will be in the near future.
Strategic factors • Desktop computing power • hardware • fastest affordable processor • largest affordable hard drive • largest affordable monitor • removable media drives support
Strategic factors • Desktop computing power • the software it should support • metadata tools • statistical software • network tools • compression utilities
Strategic factors • Large quantities of disk space on a fast system • uncompressing files • unpacking files • package and compress files
Strategic factors • Access to at least one fast processing machine with statistical software • powerful Unix workstation • handle larger-scale problems
Strategic factors • Mass storage that supports multiple-user access to files and preferably multiple-system access • distributed file system • institutional repository
Strategic factors • Support software for data services • statistical packages • metadata support • communication tools • web tools • blogs/wikis • training tools
Strategic factors • Network connectivity permitting high-speed file transfers • off and on campus transfer • may require using services elsewhere on campus
Implementation strategy • System administration takes a lot of time! • think twice about becoming your own system administrator
Implementation strategy • Purchase compatible computing equipment • to receive maintenance support • simplifies the sharing of peripheral devices
Implementation strategy • Investigate local computing support • possibly a centralized high performance computing service or compute grid • site licenses for software
Implementation strategy • Align with local institutional repository services and digital preservation initiatives • introduce data to the planning of your institutional repository
Data infrastructure models • Data centres • The data centre is part of the instrumentation infrastructure. • e.g., the Large Hadron Collider • Data repositories • The repository is part of a specific institution’s larger stewardship mandate for digital resources. • e.g., Odesi in Scholar’s Portal • Domain archives • Domain archives are institutions established explicitly to preserve and provide access to a domain’s data. • e.g., the ICPSR and the UKDA
health physics soc sci biology astro clinical data LHC data biology data astronomy data scientific data infrastructure computing/data grid infrastructure GÉANT network infrastructure Data as an infrastructure ICPSR data Source: Mário Campolargo Open Grid Forum Barcelona, 3 June 2008 source: eSciDR study (adapted)
Information Authenticity Quality Longevity Collections: data, work-flows, publications, learning materials, etc. Repositories services Ease of use Availability Reliability Deposit, annotation, delivery, visualisation, search, help, etc Trusted Open Well managed Repositories Repository management, curation, physical security, etc Standardised Stable Flexible Access Authentication, authorisation, logical security, federation, portals, etc Transparent Responsive Informed Management Grids, Virtual Organisations, etc Physical infrastructure Available Scaleable Reliable Networks, computing, HPC, physical storage, etc e-Infrastructure for repositories e-Infrastructure of repositories e-Infrastructure for repositories Source: Mário Campolargo Open Grid Forum Barcelona, 3 June 2008 source: eSciDR study (adapted)