Read a sample: 1. Define the Move Before Choosing the Tools
A migration can begin with a reasonable technical question: Which Azure service should run this application?
It is rarely the first question worth answering.
Before selecting a service, a migration tool, or a target architecture, the team needs to know what the move is meant to accomplish. A business may want to leave a data center, improve recovery, reduce operational effort, support growth, or address aging infrastructure. Those goals sound compatible until a decision forces a trade-off. Leaving a data center quickly might favor moving an application with minimal changes. Reducing operational effort might favor changing the application or its supporting platform first. Improving recovery might require work that does not fit either schedule.
Without agreed priorities, the team can produce an efficient migration that solves the wrong problem.
This chapter establishes the planning foundation for the rest of the book. You will define a migration objective, identify constraints and stakeholders, distinguish evidence from assumptions, and record decisions in a form the team can revisit. You will also meet a fictional business application that will accompany us through inventory, architecture, cost estimation, testing, cutover, and stabilization.
The aim is not to eliminate uncertainty before work begins. It is to make uncertainty visible enough that the team can decide what to investigate and what not to promise.
A migration objective describes a desired business outcome, not merely a technical destination.
“Move the servers to Azure” identifies an activity. It does not explain why the servers are moving, what must improve, or what the business is willing to change. A stronger objective connects the move to a need and leaves room to evaluate different approaches:
> Move the order-management workload out of the current data center while preserving agreed order-processing capabilities, meeting recovery requirements, and giving the operations team a supportable environment.
That statement still needs measurable criteria and an agreed deadline. Its value is that it gives the team a basis for discussion. If someone proposes a design that reduces server administration but disrupts order processing, the team can evaluate it against the objective rather than debating technology in isolation.
A useful migration objective answers four questions:
- What is changing? Identify the workload or business capability in scope.
- Why is it changing? Describe the business pressure or opportunity.
- What must remain true? Name essential service, security, data, and operational requirements.
- How will the outcome be judged? Identify the people who will accept the result and the evidence they need.
Keep objectives separate from preferred solutions. “Use a managed database” may eventually be a sound decision, but it is not the same as “reduce the effort required to maintain the database.” The first commits to an approach; the second describes an outcome against which several approaches can be compared.
Some constraints will be fixed by outside circumstances. Others may only appear fixed because no one has asked the person with authority to change them. Record the difference. A contractual data-center exit date, for example, deserves a different treatment from an internal target that a sponsor can renegotiate.
Throughout this book, we will plan the migration of Northstar Order Desk, a fictional internal application used by a fictional distributor, Northstar Supply.
Order Desk allows employees to enter customer orders, check their status, and produce a daily fulfillment file for a separate warehouse system. The application currently runs in a company-operated data center. Users access it through a web browser, and the application stores order information in a database.
The initial project discussion has produced the following picture:
- The business wants to reduce its dependence on the current data center.
- Order Desk is important to daily operations, particularly during order-entry hours.
- The application appears to rely on company identity services, internal name resolution, and a scheduled exchange with the warehouse system.
- An infrastructure team maintains the servers, while an application support team handles application incidents.
- The finance team wants a credible view of both migration spending and ongoing operating costs.
This is enough to begin a conversation. It is not yet an inventory or a dependency map. We do not know, for example, whether the warehouse exchange runs in only one direction, how the application authenticates its service accounts, or whether the current recovery procedures meet business expectations.
For the example, Northstar’s sponsor proposes this objective:
> Relocate Order Desk from the current data center to an Azure-hosted environment, without losing accepted orders, while maintaining agreed business operating hours and establishing a supportable recovery process.
The statement is deliberately broader than a server move. It includes the application, its data, the people who operate it, and the business process it supports. It also contains terms that need clarification. “Without losing accepted orders” requires an agreed definition of an accepted order and a way to verify it. “Agreed business operating hours” requires confirmation from the people who use and support the application. A “supportable recovery process” requires both a target and a tested procedure.
Those clarifications are planning work, not reasons to delay the first draft. A draft objective gives stakeholders something concrete to review.
A workload boundary defines what the migration team is responsible for moving or changing. It also identifies external systems that must continue to work even if they are not part of the migration.
For Order Desk, a first boundary might include the web application, its database, application configuration, scheduled jobs, and…