Co-Location Guide for Managed Facilities
Summary
ITS is the custodian of the computer center and acts as the responsible agent on behalf of the Department of Electrical and Computer Engineering. This document provides guidance for both ITS staff and potential co-location tenants with regard to governance of the space.
Objectives
Create a space for ITS to house a minimal amount of operational disaster recovery systems
Create a space for ECE researchers to house their computer systems which are vital to achieving their desired outcomes but for which no other viable co-location options exist
Reduce the amount of server class systems housed in areas unsuitable for their operation
Provide record keeping of tenant systems to allow for continuity of ownership and visibility into computing resource usage
Scope
Usage of the space and the systems housed within it shall adhere to these requirements
Rack mounted or tower servers actively used for research shall be permitted to be located in the managed facilities
Tenants under consideration shall be members of the Department of Electrical and Computer Engineering and have a clear research purpose for ECE
Qualification for admission of tenant systems shall be judged by the capacity of the room, the condition of the equipment, and the overall accountability of the owning tenant (typically an ECE Faculty member)
All tenant candidates and their equipment shall be screened by ECE ITS
Unidentified equipment shall be disconnected and removed from the space at the cost of the owner
ECE ITS will only setup equipment in the presence of the owner
Access to the room shall only be granted to designated system maintainers during standard University business hours
Tenants in the room shall be accompanied by an ECE ITS staff member or designated representative at all times
Tenant equipment shall be placed in an area designated by ECE ITS with consideration for all other tenants
The space shall not be used for short-term or long-term storage of inactive, unused, or spare equipment
Movement of equipment, power, or network connections without prior approval and supervision of ECE ITS is prohibited
ECE ITS Responsibilities
ECE ITS in no way assumes responsibility for hardware or software maintenance of tenant systems
ECE ITS will ensure the proper functioning of the room environment including power, networking, and HVAC from the upstream provider down to ITS owned and operated equipment only. Operation and functionality of equipment between ITS’ equipment and the tenant’s equipment is the tenant’s responsibility entirely.
ECE ITS assumes no responsibility for damage or diminished functionality of tenant systems as a result of changes to the room environment
ECE ITS reserves the ability to shut down and disconnect any system in the room in emergency situations to prevent harm to ITS and/or tenant systems
Planned maintenance that affects the functioning of room services upon which tenant systems depend shall be communicated to the tenant contact of record in advance with sufficient notice given the circumstance
Video surveillance will be installed and operated by ECE ITS as an additional way to monitor room conditions. Audio will not be monitored or recorded.
Tenant Responsibilities
Submit a record of all equipment housed in the room upon first occupancy and at least once every 6 months. See addendum for an example equipment update message.
Affix an adhesive pouch to the equipment where possible that includes basic information about the system and contact information. ITS will provide the pouch.
Respond to ECE ITS inquiries about room equipment in a timely fashion. If equipment contacts are unresponsive to ECE ITS inquiries for more than one month, the respective owner’s equipment will be disconnected
All equipment and associated costs for maintenance shall be paid for by the owner, including mounting hardware. ECE and/or ITS shall not pay for nor subsidize the cost of installing, housing, mounting, or maintaining the hardware.
Respect other co-location tenant’s equipment and make every attempt to avoid disrupting their operation
Keep computing systems and associated equipment in working order such that other room systems are not adversely affected. This includes but is not limited to aspects of computing such as network bandwidth consumption, power draw, heat dissipation, cable management, cable safety (sheathing, insulators, etc.)
Notify ECE ITS immediately of any changes in equipment state via the official ticketing system - help+colo@ece.cmu.edu
Notify ECE ITS immediately of any ownership or contact changes
Equipment Update Message Example
Upon occupancy and every 6 months thereafter starting from the date of the equipment owner’s engagement with ECE ITS, owners shall submit an update to ECE ITS including contacts for the equipment, equipment names, serial numbers, purchase date, any significant changes in system internals, any significant changes in external power or network connectivity.
From: john.owner@andrew.cmu.edu
cc: julie.contact1@andrew.cmu.edu, jessie.contact2@andrew.cmu.edu
Subject: Bi-annual colo equipment update
Item 1: Dell PowerEdge R720
Form factor: 4U rackmount
Serial: XB7T4MD
Manuf. Date: 1/3/2018
Hostname: goodserver.andrew.cmu.edu
HW address: 00B24CA19046
OS: Ubuntu 16.08
HW characteristics: GPU server with 2 full-height, full-length x16 PCIe adapters, 2x2 PSU @ 1500KWh ea, single network interface
Primary users: ngordon@andrew.cmu.edu, jbterisa@andrew.cmu.edu, qbertova@andrew.cmu.edu
Item 2: Dell PowerEdge R720
Form factor: Full height tower, extra wide
Serial: XB7T4MF
Manuf. Date: 1/3/2018
Hostname: betterserver.andrew.cmu.edu
HW address: 00B24CA19042
OS: RHEL 7.6
HW characteristics: 4 CPU and high-memory compute cluster for Matlab and Comsol simulations
Primary users: ngordon@andrew.cmu.edu, jbterisa@andrew.cmu.edu, qbertova@andrew.cmu.edu