Cloud Migration

The What

The district goals will determine what hardware, platform, and software resources will move to the cloud. If the goal is cost savings, it is important to do a cost analysis before deciding what resources to move. If the goal is improved user experience or simplification, it is critical to take into account how much of the district resources need to be included to achieve critical mass for those improvements.

Some resources may not be cost effective to move to the cloud. A brand new data center may meet most of the district’s current needs and the cost of that facility and hardware would essentially be a loss if everything was moved to the cloud. Some software systems are simply not architected to take advantage of cloud with respect to parallel execution and load sharing. In such cases, the district may be better off keeping some resources in a local data center, perhaps improving resource usage via virtualization.

The When

Moving to cloud computing is not an overnight process. Different cloud services may make sense to utilize over a timeline of months or even years. For example, a first step might be to begin using cloud-based collaborative or content creation tools. A next step might be to move some systems to the cloud. A final step might be to go all-in on cloud by moving as many data center resources as possible as quickly as possible, eventually making the data center obsolete.

When to look at moving to the cloud:

  • When servers are at their end of life, move those workloads to the cloud
  • If storage capacity is limited, shift data to the cloud
  • Take large data sets from on-prem centers and analyze in the cloud
  • Do app development and testing in the cloud
  • Put student computer science and web development into separate cloud accounts
  • When selling district buildings, don’t move servers, move to the cloud

The Why

Before embarking on a cloud migration, it is critical to know the goals of that migration. Rationales that boil down to “cloud is good” are inadequate. Consider whether your district is looking for cost savings or improved user experience or agility in trying out new products or a foundation for simplified IT management or something else. The goal(s) of the district will drive the decision making regarding what systems can move to the cloud.

The How

When embarking on a cloud migration, always:

  • Develop a migration plan that addresses the goals of the district; the trade-off analysis of the benefits based on each of the district goals; the overall policy for migration; the timeline for migrating each resource; and what resources to keep locally,
  • Develop the security policies and governance to support the district security responsibilities.
  • Write a service level agreement that passes district policies onto the Cloud Service Provider and their vendors and sub-contractors; clearly spelling out the responsibilities of the district vs. the CSP, and providing guarantees and auditing of security for both physical access and the computing environment.

Best Practices

Best Practices to achieve cloud benefits include:

  • Identify price differences for hosting in different regions;
  • Ensure data is hosted close enough to support latency requirements, yet far enough (or with sufficient redundancy) to support disaster recovery;
  • Ensure data is hosted within the United States;
  • Beware of translating existing data centers directly to the cloud as this could leave to overspending, and;
  • Expect CSPs to identify areas where the district cloud could be more secure and look across all district cloud services to give recommendations on opportunities to save money.

“There isn’t a singular reason districts should consider a move to the cloud. Rather, there are a host of them. Districts should be considering a move to the cloud because:

  • It makes technology more flexible, in both capacity and dollars, to rapidly changing demands for services.
  • It engages technology staff in a conversation about the true total cost of running technology on site, including buying hardware, cooling and  powering equipment, and staff time for setup and repair. It then shifts the traditional IT budgeting model to a utility model where you pay only for what you use and, more importantly, you don’t pay for what you don’t use (like electrical power).
  • It frees time for technology staff to better support applications and their users.
  • It delivers an easy opportunity for technology departments to deliver resilient services that are nationally or even globally dispersed.
  • It removes some liability for information security.
  • It’s the future of IT.”

—Stuart Long, CIO, Clackamas Educaton Service District