ARCHIVED: At IU, what options does UITS offer for purchasing computing resources in the UITS secure Data Centers?

This content has been archived, and is no longer maintained by Indiana University. Information here may no longer be accurate, and links may no longer be available or reliable.

The following options are available through UITS for the purchase of computer resources in the Data Center:

  • Condominium computing (UITS-managed, departmentally-owned research IT systems or system components): "Condominium cluster" environments allow researchers, research labs, departments, and schools to have computational nodes housed within the IU Bloomington Data Center and managed by UITS Research Technologies staff.

    Departmentally owned systems

    If you or your IU sub-unit have money available to use for a computing cluster, you always have the option of buying and operating your own cluster, as well as managing it, backing it up, and securing it against hackers. For departmentally owned clusters housed in the IU Bloomington Data Center, UITS will also provide hosting services under a facilities management agreement. Such hosting agreements improve physical security and reliability of power supplies to clusters. UITS offers these hosting services for a fee that reflects the cost of Data Center floor space, electricity, and cooling.

    To inquire about a departmentally owned cluster in the Data Center, email dcops@iu.edu.

    UITS-managed systems

    Alternatively, you can purchase nodes that are compatible with IU's Carbonate cluster, have them installed in the very secure IU Bloomington Data Center, have them available when you want to use them, and have them managed and secured by UITS Research Technologies staff. This colocation service gives schools and departments access to compute nodes that are dedicated solely to their use within Carbonate's physical, electrical, and network framework while leveraging the security and energy efficiency benefits provided by location within the IU Data Center.

    To inquire about the colocation service on Carbonate, contact the High Performance Systems (HPS) team.

    The typical characteristics of a cluster condominium hosting agreement are:

    • The type of resource included within the cluster condominium supports uni-processor or scalable parallel jobs (e.g., Condor, MPI, or OpenMPI), and the nodes are compatible with the basic infrastructure of the Karst cluster.
    • Schools, departments, labs, or individual investigators will have and irrevocably retain ownership of the equipment they purchase, which is housed within the cluster condominium service. However, the lifespan of equipment will rarely exceed four years, and never exceed five years, under UITS management via a Memorandum of Understanding (MOU).
    • Job scheduling is based on node owner affinity to ensure fairness. When the owner of a certain pool of hardware wants to use that hardware, the owner's jobs are prioritized.

    It is also possible to purchase storage space within a variety of disk-based storage systems, including the Research File System, ARCHIVED: Data Capacitor II, and the Scholarly Data Archive.

    For departmentally-owned clusters, UITS also will provide hosting services under a facilities management agreement. Such hosting agreements improve physical security and reliability of power supplies to clusters. Since they do not have the same community benefit aspects that condominium clusters have, UITS offers them at a cost that reflects the cost of Data Center floor space, electricity, and cooling. Nodes in the IU condominium facility are monitored electronically and UITS staff 24 hours a day, 365 days a year. If you want to own your own computing resources, UITS can help you do that securely and safely.

  • VM hosting for research:
    • VMs hosted in XSEDE: If you have a web portal, web accessible service, or web-accessible database used by research labs outside of IU, or that you intend to be usable beyond the IU community, you can apply for and receive an allocation of virtual machines and storage space on the federally-funded Extreme Science and Engineering Discovery Environment (XSEDE) service. IU operates this service under a subcontract from the National Center for Supercomputing Applications (NCSA).

      The Quarry Gateway Web Services Hosting System comprises multiple Dell AMD systems geographically distributed for failover. Each system has at least eight cores and 32 GB of memory. Persistent storage is available via local home directories or the Data Capacitor Wide Area Network (ARCHIVED: DC-WAN) Lustre file system. The system uses OpenVZ to provide virtual hosting of RPM-based Linux distributions. The host operating system is Red Hat Enterprise Linux. Virtual machines hosted on the system are backed up daily using the IBM Tivoli Storage Manager (TSM).

      Several IU web services and science gateways already run in this service (see Gateway Listing). You have to apply for an allocation through XSEDE to use these VMs, but they are otherwise free. If you are interested, contact Research Technologies for assistance with the XSEDE allocation process.

    • II-Research: II-Enterprise is engineered for services that demand a very reliable and robust computer environment, such as mission critical services. That is more reliability than is needed for many scientific applications. UITS is planning II-Research, a third variant of Intelligent Infrastructure.

      II-Research will be based exclusively on tools consistent with national best practices in research, using open source tools whenever possible, with ties to UITS Research Technology storage facilities such as Data Capacitor II and the successor to the current Research File System. This will be a chargeback service, as it replaces services that have been based predominantly in departments and post-dates the allocation formula that supports the baseline services offered by Research Technologies, previously funded from departmental funds. It will be priced at a cost below II-Enterprise.

This is document bdqm in the Knowledge Base.
Last modified on 2018-01-18 17:22:48.