Server Operations Managed Servers Service Level Agreement (SLA)
The goal of this SLA is to delineate responsibilities, communication paths and other details for computing capacity and support provided by IS&T Server Operations. This agreement covers: Client: Server Name: Effective Date: Annual Cost:
Client Contact Information
Business Contact - Primary: Business Contact - Secondary: Group Email address: Technical Contact - Primary: Technical Contact - Secondary: Financial contact: Account number for service charges: For pageable servers: Emergency Contact - Primary: Emergency Contact - Secondary:
IS&T Server Operations Representative Departmental Financial Contact
Departmental Technical Contact
We will install, house and maintain the machine, including appropriate software, in an access controlled machine room that has UPS power, with a generator backup for servers in our primary facility. We will also provide the machine with network drops and hostnames/IP addresses. We will provide system monitoring and provide support based on the Service Level described below.
SERVER SPECIFIC INFORMATION AND CHARGES
Cost Factor U factor OS/Hardware Age of Server Service Level Storage Backup
Total Points Total Lease Total Charges Description of server purpose and/or any special services:
Basic Cost Factor Components
The following four costs apply to all servers managed by Server Operations, both physical and virtual: Supported or Unsupported OS/Hardware 1 point Supported 2 points Unsupported Level of Service 1 point - 9x5 2 points - 24x7 4 points - Pageable Storage 0 points for internal storage 1 point per 100GB for external mid-range storage 2 points per 100GB for external high-end storage Type of Backup 1 point for TSM 2 points for NetBackup or other custom backup solution. The two costs below apply only to physical servers and will not be billed to virtual servers: Size of Server (U factor / 1.75") .5 point per U Age of Server 0 points year 1-4 1 point per year 5 and over Note: A "point" is currently assigned a value of $1,100 as of MIT FY09.
SERVER OPERATIONS POLICIES AND PRACTICES
The section below describes specific expectations of Server Operations and the Business Owner. Of special note are items in italics and underlines.
Billing will be on a quarterly basis. The SLA may be altered or cancelled on a quarterly basis. SLA modifications that result in changes to cost will be reflected in the subsequent quarter.
Hardware supplied by Server Operations will by default meet the following minimum standards: - dual hot swappable power supplies - mirrored disks for the OS - redundant cooling fans - redundant paths to external storage - common components kept on site After initial configuration, significant changes to the system will be discussed either via email or in a meeting prior to being implemented, and will be reflected in future costs. Most servers are replaced when they are four years old, but this may vary depending on specific circumstances. If this SLA is renewed beyond the life of the hardware of the machine, we will recommend new hardware configurations. Actual charges are determined by the point system described at http://web.mit.edu/ist/services/hardware/serverops-pricing.html.
OS and other Software
We will maintain the operating system on the machine. We will also maintain the "standard" Server Operations environment on the machine, including remote login (ssh and other means) and required root access to the machine via Kerberos. We will coordinate fixes necessary to address security and other support issues for the server. IF REQUESTED, we will also maintain a "standard MIT" Apache web server installation (excluding web content) on the machine. In all of the above, "maintain" will include initial installation and the installation of: any subsequent upgrades, security, and other necessary patches. When a planned outage is required to perform preventative maintenance on part of the system that we maintain, we will coordinate a time with you. Also available, as requested, are Oracle database services that include: installation, monitoring, and ongoing patches and upgrades. We will evaluate and recommend specific Oracle versions and configurations depending on application requirements.
While all servers are monitored 24x7, the service level determines what action is taken prior to the next business day. The three levels of service are: 9x5 We carry a 9x5 maintenance contract on the server and have 9x5 auto-paging set up. If paged during business hours, we will coordinate any maintenance, reboots, etc., with the business contact. Outside of business hours email will be sent to the business contacts informing them of the problem/resolution.
24x7 We carry a 24x7 maintenance contract on the server and have auto-paging set up. If paged during business hours, we will coordinate any maintenance, reboots, etc., with the business contact. If paged after business hours, we will perform emergency maintenance, reboot, etc., as deemed necessary. Email will be sent to the business contacts informing them of the problem/resolution. Pageable Emergency business contact information is mandatory. We carry a 24x7 maintenance contract on the server and have auto-paging set up. We will make every effort to reach the emergency business contact immediately upon discovery of a problem and, if possible, will coordinate any maintenance, reboots, etc., with the business contact. If the emergency contact can not be reached, we will perform emergency maintenance, reboots, etc.
We will perform regular backups of the machine. This includes daily system backups including Oracle, when applicable. In special cases, our standard backup procedures can be modified to meet specific needs. The default retention for daily backups is 10 weeks. Off-campus storage of data is an available option.
ServerOperations provides client application monitoring, including automated paging, which will be coordinated with the client. For accurate monitoring of performance and to plan maintenance, it is expected that the client will supply Server Operations with basic application usage statistics by time of day and day of week so that we can better monitor performance and plan maintenance. This is especially important for non-business hour maintenance when we must coordinate activities across multiple machines.
We will agree upon established maintenance windows when a service interruption may be necessary. These are typically Tuesday and/or Thursday 6-9 pm and occasional weekend outages. We will make every effort to coordinate all maintenance outages with the client, but there may be cases where emergency intervention is necessary that may result in an unexpected interruption of service. The client will identify the times of year when availability of the server is most critical. Server Operations will use this information for advance planning of maintenance activities, especially where multiple systems may be impacted.
We will work with you to establish a demonstrable Disaster Recovery / Business Continuity plan appropriate for the Business Category of the supported application. This may require your participation in Disaster Recovery exercises to another location
Contacting IS&T Server Operations
For all requests, you may contact us by sending mail to ASST-HELP@MIT.EDU In the event of an operational emergency, please use the following procedures to contact IS&T Server Operations:
call 617-324-2778 (4-ASST) - leave voice mail stating: - the impacted service / name of server - the nature of the problem - your name and contact phone number if you are not contacted within 30 minutes - call 617-324-9080 - leave voice mail indicating this is a follow up call stating: - the impacted service / name of server - the nature of the problem - your name and contact phone number