Hybrid cloud service-integration partnership for a large public research university

The situation

A large U.S. public research university wanted a hybrid-cloud strategy, not a public-cloud migration. The constraints were specific: academic-calendar release windows, research-computing workloads with their own scaling profiles, federated identity that already spanned dozens of campus systems, and a procurement environment that required commercial neutrality between cloud providers.

They had already evaluated several proposals oriented around a single hyperscaler. What they wanted instead was a service-integration partner who would design the operating model first, then run a migration factory that could move workloads across both public and private cloud as the architecture demanded, not as a single vendor's default would dictate.

What we did

1. Hybrid Cloud Service-Integration Partner SOW. Multi-phase Statement of Work covering assessment, design, pilot, and phased migration. Phase 1 (Jan 2017 – Sep 2017): assessment, design, pilot. Phase 2 (Oct 2017 – May 2020): migration of approved waves with testing and certification gates. The structure was designed for academic-calendar coexistence.
2. Migration factory operating model. Designed the migration factory: discovery, assessment, application remediation, application testing, deploy. Built the model around repeatable workstreams that could process the application portfolio in waves, each with its own pilot-application selection, remediation, and certification cycle.
3. Per-application migration runbook. Application-owner provides existing test strategy, documents, and test cases. Migration factory team conducts and supports unit, integration, regression, performance, and UAT cycles, certifying each migrated application before cutover. Target environment provisioning via configuration management automation.
4. Cloud management platform and operating-model selection. Worked across both public and private cloud providers to select the right cloud management platform for the hybrid environment, with the operating model that the university's IT team could run after the migration factory engagement closed.

Outcomes

• Multi-year hybrid-cloud program structured into formal phase 1 (assess + design + pilot) and phase 2 (migrate)
• Migration factory designed and stood up with repeatable per-application workflow
• Cloud management platform selected with vendor-neutral evaluation across public and private cloud providers
• Operating model designed for hand-off to the in-house IT team at the end of the engagement

What was hard

Higher-ed migration windows are unforgiving. Spring break, end of semester, and the summer maintenance window are nearly the only times you can take a system down. The factory had to be sized to move waves of applications between those windows, which meant running parallel workstreams in a way that smaller engagements would not require. Sequencing was the hardest design decision.

The other hard part was federated identity. The university already had a multi-system identity environment that no migration plan could safely simplify. We spent more time than we expected on identity-flow mapping for each application and on pilot validation that did not break shibboleth federation. The migration approach had to be informed by the identity architecture, not the other way around.

Selected artifacts

• Hybrid cloud SOW structure. Multi-phase Statement of Work for assessment + design + migration
• Migration factory operating model. Discovery → Design → Pilot → Migrate with per-wave testing gates
• Per-application migration runbook. Application-owner test strategy + factory testing cycles + cutover playbook
• Cloud management platform evaluation rubric. Vendor-neutral evaluation across public + private cloud providers

Public-cleansed templates and frameworks drawn from engagements like this one are in our resources library.