Request a Server.
Fill out the survey for a new Virtual Server
Once a request is received, the server group will identify a two day period to review and approve the request. A member of the system team will contact the original requestor to approve the server quote, and then begin progress to build the server.
About our Servers
Server location will be determined by optimum resource allocation. Tufts has two datacenters available in our virtualization environment, one located in the the Somerville Tufts Administrative Building (TAB) and a second on Summer St. in Boston, MA
- Datacenter location and Location Independence for the service is determined by our Service Classification below.
- Our expertise and troubleshooting is gratis as part of the service, but depending on your storage and server requirements some costs may be associated with hosting your service. Before a server is built the extent of the cost model will be discussed with you and approved via IDR.
- Our servers are virtual hosts and are either RedHat Linux or Windows Server. It may not be possible to support requests of operating systems outside of our standard, although on a case by case basis we can make allowances for self-enclosed Appliance Images.
- Servers are backed up digitally in accordance to our back-up retention policies. Server back-ups are made available for disaster recovery reasons only.
- We monitor our enterprise environment using a number of monitoring systems and support many basic service health checks.
- Hosted servers are patched in the months of Dec-Jan and Jun-July each year. Patching will require downtime per host, but depending on how the application/service is architected may not require a service disruption.
- Standard maintenance windows are below as well as Uptime targets.
Uptime
TTS strives to maintain high service availability for all of our services. In order to ensure we are targeting the correct uptime models, we will classify your service based on the following criteria:
- The ability of the application/service to target high-availability through design and implementation.
- Impact of your service to the core missions of the University.
The following is a matrix we use once a service has been classified
Service Class |
Expected Behavior |
Type of Service |
|---|---|---|
Class 1 |
Services are expected to run primarily from our Tier IV Datacenters but be able to automatically run from our Tier I Datacenters in order to maintain high availability |
Mission Critical Services for the University that directly impact the core business on a 24/7 level. |
Class 2 |
Services are expected to run primarily from our Tier IV Datacenters but will not automatically transition from this location to a Tier I location in the event of a service interruption. It may be possible to regain service to Class 2 applications should the Tier IV datacenter be unavailable but only after Class 1 applications are successfully running |
Mission Important, this service may be accessed on a 24/7 basis but does not critically impact the core mission of Tufts |
Class 3 |
Services are expected to run primarily from our Tier I Datacenters and will not be easily able to transition to a Tier IV datacenter in the event of a service interruption. |
Business Important, this service may be important during the expected "business hours" of the university but does not directly or indirectly impact the core missions here at Tufts. E.G print servers |
Class 4 |
Services are expected to only be available at a limited basis, or not critical if unavailable outside of university "business hours." |
Business services such as test or development services that are not directly public-facing |
Maintenance
Planned vs Unplanned
Planned maintenance is discussed, tested, and scheduled prior to 48 hours before the occurring maintenance. For servers classified as "non-production" it may be necessary to schedule and complete work outside of the 48-hour prior window or outside described maintenance windows below.
Schedule for Planned Maintenance Windows
MORNING |
USED FOR |
DATES |
|---|---|---|
6am-8am MWF |
No Impact, or No major Rollback required |
Monday through Friday |
6am-10am |
Potential for larget impact, requires elongated window, or poor rollback options |
SUNDAY |
EVENING |
USED FOR |
DATES |
12am-2am |
Outside of business hours, helpful to have work on Friday |
FRIDAY |
9pm - 11pm TuTh |
Low Impact, or No major rollback required |
Monday - Friday |
Accessing your server
We will provide appropriate level access to the server for approved administrators. Shell and Remote Desktop access will require access to the Tufts VPN.
Communication
When creating servers for you, we will collect a primary service owner and e-mail contact for your team. If this communication entry should change, please update us of this. For any planned maintenance or changes to our services we will communicate via this collected communication.
In the event of an unplanned interruption to your service, EI will communicate through one of the following communication vectors:
- Always
- Service Owner escalation if identified as required, or for Class I or II services by default.
- Sometimes
- Tufts Service Advisory Team Discussion List (SAT@elist.tufts.edu)
- Tufts Front-line Service Personnel Discussion List (FSP@elist.tufts.edu)
- Tufts Infoboard (infoboard.uit.tufts.edu) if a service or services are being impacted requires direct customer communication.
- Tufts Service Announcements Discussion List (tied to Infoboard Postings)
Costs
Servers as described above are provisioned at a $5,000 per four year cost basis, per server. For a service with a single production and single test virtual machine, the cost is calculated as two servers, or $10,000 per four year cost basis. This cost is to be billed by Inter-Departmental-Requisition once the server is provided to the customer.
There are a few things that can modify the cost of the server:
- Servers that require a SQL Server Database must also include the price of the license for the database which is dependent upon the number of processor cores in the virtual machine
- Servers which fall into the "other" category for RAM and Processor requirements may incur additional cost, to be discussed with the responsible server provisioning group.
- Servers which have large storage requirements that exceed 100gb of enterprise storage may incur additional cost, to be discussed with the Enterprise System and Services department.
Resource Allocations (CPU, RAM, Storage)
- We will track performance usage of provisioned servers and in the cases where server requests were over-provisioned, we may recommend a course of action to reduce resource allocation or virtual location of the server. In most cases a simple reboot will allow for these changes to take place during a defined maintenance window.