IT Services
Service Level Agreements Image

High Performance Computing
Service Level Agreement 2016/17

Quicklinks

  1. General Overview
  2. Service Description
  3. Roles and Responsibilities
  4. Requesting Service
  5. Hours of Coverage, Response Times and Escalation
  6. Maintenance and Service Changes
  7. Pricing
  8. Review and Reporting
  9. Approvals


1 General Overview

This is a Service Based Agreement (SBA) between the campus community and the Division of Information Technology to document:

  • The suite of High Performance computing services provided by the Division
  • The general levels of response, availability, and maintenance associated with these services
  • The responsibilities of the Division as a provider of these services
  • The responsibilities of the clients receiving these services
  • Departmental services and processes not documented in the IT and Campus SLA

This Agreement is valid from April 2016. Review is every two years, with significant input and guidance provided by the University Advisory Committee for Research Computing, or as otherwise needed.

This document describes the IT division support model for computational computing clusters. This document is intended to outline the expectations and limitations for the use of, access to, and support of dedicated high performance computing services and resources for research.


2 Service Description

The description of the suite of High Performance Computing services provided is on the Division services web page at: http://DoIT.creighton.edu

2.1 Service Scope

DoIT provides short and long term data processing in support of Creighton related scholarly research projects, programs, education, and operations. The High Performance Computing (HPC) team provides infrastructure, people and processes to support the following suite of services.

HPC clusters require a significant level of expertise to build, operate and maintain. There are a high number of potential failure points inherent to clustered computing systems.  This SBA describes a collaborative research community required to provide these complex services.  Leveraging the economy of scale and experience allows Creighton Univeristy to create the capacity to offer valuable services for a minimal investment as compared to fully commercial research computing offerings. Under advisement from the faculty, the University Advisory Committee on Research Computing, and the greater research community, standards are applied to create a repeatable implementation model.  This model will allow for scalability from the perspective of the administrator and the researcher to create customized computing environments to meet and exceed the diverse research needs of Creighton University.

Please refer to the High Performance Computing link above for additional information about these services:

Active Directory (Campus AD)

  • File storage and sharing
  • Authentication
  • Physical Security
  • Environmental stability

Data Center services – See Data Center SLA for a description of services provided.

Hosting

  • Physical Server Hosting
  • Virtual Server Hosting (preferred over physical)
  • Coordination of cloud hosting.  Example:  Oracle HRIS
  • 24x7 Data Center Operations *HPC support is limited to systems in the production environment only. All experimental and developmental HPC systems are unsupported.
  • Shibboleth authentication

Managed Services (including limited server administration)

  • UNIX Systems
  • VMware
  • Windows Systems
  • Storage

HPC Services

  • Basic Scientific Linux installation on the master node
  • Installation of licensed or open source software utilized for data processing and visualization
  • Various cluster implementation toolkits
  • Shared, Federated scheduler (SLURM, ROCKS or a equivalent standard management and scheduling application otherwise recommended by the Univeristy Research Computing Advisory Committee)
  • Purchase and procurement consulting and support
  • Initial cluster build and setup
  • Limited cluster debugging and testing support
  • Operating maintenance
  • Training on scheduling and accessing the cluster
  • System and network monitoring
  • Hardware monitoring
  • Environmental control for systems within the University Data Center
  • Warrantied repair and replacement of faulty hardware and software
  • NFS storage for servers managed by the IT Systems Administration team
  • Storage for students, faculty, and staff, accessible through the UNIX timeshare or DFS.

In addition, the Division provides infrastructure, people, and processes to support:

  • Service management to facilitate High Performance Computing and planned service expansion/improvement
  • Limited support services during Univeristy business hours

2.2 Service Level Performance

For support requests submitted through the EasyVista system, first-response times and resolution times are tracked and reported as part of the IT and Campus SLA. The report is posted quarterly in the IT Service Catalog, under Service & System Metrics.


3 Roles and Responsibilities

3.1 Parties

The following Service Owner(s) will be used as the basis of the Agreement and represent the primary stakeholders associated with this SLA:

Stakeholder Title/Role   Contact Information
Ryan Cameron Research Computing Group Chair Advisory ryancameron@creighton.edu
Tim Brooks VP   timbrooks@creighton.edu

        
3.2 IT Responsibilities

The Division will provide the infrastructure, technology, people, processes and monitoring tools necessary for High Performance Computing and:

  • Clearly document services provided in the IT Service Catalog.
  • Meet response times associated with the priority assigned to incidents and service requests.  See “DoIT and Campus SLA” for details.
  • Generate quarterly reports on service level performance.
  • Provide appropriate notification to the campus community for all scheduled maintenance via the Maintenance Calendar, Service Catalog web page and/or a communication to campus via the IT Communication Lead.


3.3 Customer Responsibilities

Client responsibilities and/or requirements in support of this Agreement include:

  • Utilizing the Service Desk for incidents
  • Contacting the IT Project Management Office, (via the service desk) for additions or changes in established service levels

Adhere to Creighton and IT policies; this includes:

  • Adhere to the Creighton Acceptable Use Policy
  • Adhere to the Creighton Minimum Network Connectivity Requirements Policy
  • Adhere to the DoIT Data Center Access Policies and Procedures (customers with access to the Data Center only)
  • Data Storage Policy
  • University Policies can be found at: https://www.creighton.edu/generalcounsel/cupolicies/

Additional Roles and Responsibilities for IT and the Customer are posted on the Data Center Service Pages for the following:

  • Managed Services
  • Physical Server Hosting
  • Virtual Server Hosting
  • Data Backup and Recovery on warrantied HPC equipment

4 Requesting Service

See the IT and Campus SLA for standard methods of contacting DoIT for service.

Physical and virtual hosting requests can also be made on the Division webpage.


5 Hours of Coverage, Response Times & Escalation

See “DoIT and Campus SLA

5.1 Hours of Coverage

Monitoring of the High Performance Computing cluster is provided 24 hours a day, seven days a week except for periods of planned maintenance.  Staff is on site Mon-Friday 8am-5pm, except holidays and other days the campus is closed. The High Performance Computing cluster is not a production service. Although every possible consideration has been made to ensure data has redundant back-ups and servers have processing fail-safe redundancy, the cluster is experimental. All users of the system assume some degree of risk when using the system. Researchers are strongly encouraged to consider maintaining a personal backup of all critical research data should the unexpected loss of data occur.

See “DoIT and Campus SLA” for more details.

5.2 Incidents

5.2.1 Incident Response

See the IT and Campus SLA.

5.2.2 Prioritization

See the IT and Campus SLA.

5.2.3 Service Request

See the IT and Campus SLA.

5.3 Escalation

See the IT and Campus SLA.

5.4 Information

See the IT and Campus SLA.

5.5 Other Requests

Requests for service features and functions not yet implemented can be submitted to the DoIT Service Desk though the online ticket system at doit.creighton.edu.


6 Maintenance and Service Changes

Some High Performance Computing services have pre-scheduled maintenance windows (which are superseded by campus-wide change-restricted days, as noted on the IT Maintenance Calendar). Most maintenance will not impact service.  In the event that maintenance is expected to result in a service outage, DoIT will provide at least two weeks’ notice. These pre-scheduled maintenance windows are:

  • For all associated High Performance Computing, the Division reserves a pre-scheduled maintenance window every Friday from 10PM-4AM for non-disruptive maintenance.

See the IT and Campus SLA for details of our standard maintenance and service process.


7 Pricing

The Division covers the costs relating to these services in order to encourage members of the Creighton University community to use the central computing facilities and to take advantage of the secure and power-protected environment. These include the following services:

  • Virtual (preferred) and Physical Server Hosting
  • 24x7 Data Center Operations
  • Campus Active Directory (AD) - file storage and sharing, authentication, printing
  • Shibboleth Authentication
  • Storage - for students, faculty, and staff

Cloud Computing services are arranged on behalf of the customer; costs are passed through without mark-up.


8 Reviewing and Reporting

8.1 System Performance and Availability Reporting

Support response time will be tracked and reported separately as part of the IT and Campus SLA.

8.2 SLA Reviews

The Designated Review Owner (Document Owner) is responsible for facilitating regular reviews of this document. Contents of this document may be amended as required, provided mutual agreement is obtained from the primary stakeholders and communicated to all affected parties.

The Document Owner will incorporate all subsequent revisions and obtain mutual agreements/approvals as required.

Designated Review Owner: Ryan M. Cameron, Executive Director Information Technology, Division of IT.

Previous Review Date: April 2016
Next Review Date: April 2018

This Agreement will be posted to the following location and will be made accessible to all stakeholders:

Document Location: http://DoIT.creighton.edu

9 Approvals

The Divisional Liaisons and IT Senior Managers approve this document. This document is then published on the IT Service Catalog web site along with other service based agreements. Service level information is integrated into the service page in the IT Service Catalog.