Moving to the cloud is easy to describe and hard to finish. A business decides to leave a hosting account or an on-premise server, a provider is selected, and the first applications move quickly. Then the finance system, the courier integration, the backup job and the one server nobody documented turn out to be connected to everything else.
This checklist is written for the person coordinating that work inside a Bangladeshi SME. It is organised by workstream, with an owner and evidence for each item. It does not assume a particular cloud provider.
Before the checklist: agree the outcome
Write down why the business is moving. Common reasons include ending a hardware refresh cycle, improving recovery, supporting remote staff or reducing the effort of maintaining a server room. Then write what would make the migration a failure: an outage longer than the business can tolerate, a cost increase nobody approved, or a compliance problem with customer data leaving the country. These statements shape every later decision.
1. Business ownership and constraints
- Name one accountable business sponsor and one technical lead. (Owner: managing director; evidence: written appointment.)
- List the business processes that must keep running during the migration and their acceptable outage window. (Owner: operations; evidence: a signed service window table.)
- Record data residency, regulatory and customer-contract constraints. (Owner: legal or finance; evidence: a written constraint note.)
- Agree the budget model: one-time migration, monthly run rate and a contingency for parallel running. (Owner: finance; evidence: approved budget lines.)
- Define the exit route: how data and workloads could move to another provider. (Owner: technical lead; evidence: an export and portability note.)
2. Inventory and dependencies
- List every application, server, database, file share, scheduled job, integration and domain. (Owner: IT; evidence: a complete inventory with owners.)
- Mark the criticality, peak usage, data classification and recovery target for each item. (Owner: business and IT; evidence: a criticality rating.)
- Map dependencies between systems, including authentication, DNS, email, payment gateways, courier APIs and reporting. (Owner: technical lead; evidence: a dependency diagram.)
- Identify unsupported software, end-of-life operating systems and applications with no owner. (Owner: IT; evidence: a remediation list.)
- Decide the migration pattern for each item: rehost, replatform, refactor, replace, retire or retain. (Owner: technical lead with business; evidence: a decision per application.)
3. Security and access
- Create separate cloud accounts or projects for production, staging and development. (Owner: cloud administrator; evidence: account structure diagram.)
- Enable multi-factor authentication for every administrative account and remove shared logins. (Owner: IT; evidence: an account review.)
- Define roles with least privilege and document who can change network, identity and billing settings. (Owner: security lead; evidence: a role matrix.)
- Turn on audit logging, alerting and cost alerts before the first production workload moves. (Owner: cloud administrator; evidence: a log and alert test.)
- Encrypt data at rest and in transit, and record where keys are stored and who can use them. (Owner: technical lead; evidence: configuration screenshots and a key register.)
- Review the provider's shared-responsibility model and write down which controls remain the business's responsibility. (Owner: security lead; evidence: a responsibility matrix.)
4. Landing zone and network
- Design the network range before creating servers, avoiding overlaps with office and partner networks. (Owner: network lead; evidence: an IP address plan.)
- Set up private connectivity or a managed VPN for administrative access; do not expose management ports to the internet. (Owner: network lead; evidence: a connectivity test.)
- Configure firewalls, security groups and web application protection with a default-deny approach. (Owner: security lead; evidence: a rule review.)
- Plan DNS changes, certificate renewal and email authentication records. (Owner: IT; evidence: a cutover DNS plan.)
- Confirm backup storage, retention and restore testing in the new environment. (Owner: IT; evidence: a successful restore test.)
5. Data migration
- Classify data and identify what must move, what must be archived and what should be deleted. (Owner: data owner; evidence: a data migration matrix.)
- Estimate volume, growth and transfer time using the actual link speed available to the business. (Owner: technical lead; evidence: a transfer calculation.)
- Choose a transfer method that suits the volume and sensitivity: encrypted network transfer, a provider migration service or a physical device where supported. (Owner: technical lead; evidence: a method decision note.)
- Run at least one full rehearsal and one delta rehearsal, and reconcile record counts and key totals. (Owner: data owner; evidence: a signed reconciliation report.)
- Define how data created during the final cutover will be captured and replayed. (Owner: technical lead; evidence: a freeze and delta plan.)
6. Cutover and rollback
- Write a timed runbook with named owners, commands, checkpoints and go/no-go criteria. (Owner: migration lead; evidence: an approved runbook.)
- Announce the maintenance window to customers, suppliers and staff, including how to report a problem. (Owner: communications; evidence: a sent notice.)
- Freeze changes to the affected systems for an agreed period before and during cutover. (Owner: project manager; evidence: a change freeze notice.)
- Test the rollback path, not just the forward path, and record the trigger and decision-maker. (Owner: technical lead; evidence: a rehearsal result.)
- Keep the old environment read-only and available until the new one has passed the agreed stability period. (Owner: IT; evidence: a decommission date.)
7. After migration
- Review performance, error rates, backup results, security alerts and the first full invoice. (Owner: technical lead and finance; evidence: a post-migration report.)
- Retire old servers, accounts, DNS records and licences, and remove access for departed staff. (Owner: IT; evidence: a decommission checklist.)
- Update the disaster recovery plan, asset register and support contacts. (Owner: IT; evidence: a revised plan.)
- Train staff on the new access method and support route. (Owner: operations; evidence: a training record.)
- Schedule a 30-day and 90-day review of cost, reliability and user feedback. (Owner: sponsor; evidence: meeting notes and actions.)
Common mistakes
- Migrating an application before understanding its integrations and batch jobs.
- Treating the cloud provider's security controls as a complete security programme.
- Testing the restore only after the old backup has been deleted.
- Leaving the old environment accessible with the same credentials.
- Forgetting email, DNS and certificate dependencies in the cutover plan.
- Choosing a migration pattern for every application at once instead of starting with a low-risk workload.
Metrics for the review
Track cutover duration against the plan, unplanned downtime, number of rollbacks, restore test success, monthly run rate against budget, and the number of unresolved security findings after 30 days. Add a simple user measure: support tickets per hundred users in the first month compared with the month before.
Where this fits in our work
Migration is a project, but the operating model after it is a long-term choice. If you need help with the assessment or the build, start with Custom Software & Integrations and review the engagement options on our pricing page. For the governance side of the same work, read infrastructure as code governance compared.
Sources
- Microsoft, Cloud Adoption Framework: Migrate, accessed 17 November 2026.
- Google Cloud, Migration Center documentation, accessed 17 November 2026.
- NIST SP 800-145, The NIST Definition of Cloud Computing, accessed 17 November 2026.
- FinOps Foundation, FinOps Framework, accessed 17 November 2026.
Responses (0)
No comments yet. Add the first one below.
Leave a response
Your email address will not be published. Required fields are marked *.