Colocation Migration Guide: Planning, Execution, and Risk Mitigation

Phase 1: Inventory and scoping

A complete inventory and dependency map is the foundation of a safe migration. Missing a dependency (a storage array that three production databases rely on) creates unplanned downtime during cutover.

If you are still deciding whether to renew or move, settle that before you plan the migration.

• Full hardware inventory: servers, storage, network gear, PDUs, KVMs, cabling
• Dependency map: which workloads depend on which hardware, networks, and storage
• Workload criticality classification: production, DR, dev/test, non-critical
• Network dependency map: IP ranges, VLANs, cross-connects, carrier relationships, external DNS
• Compliance and data residency constraints on what can move and when

Phase 2: Migration sequencing strategy

Sequence migrations from least to most critical. Move dev/test and non-critical workloads first to validate the process, then DR and secondary systems, then production. This gives your team practice and surfaces issues before they affect production.

For each wave, define the cutover window, the rollback trigger (what condition causes you to revert), and the rollback procedure. Never cut over without a tested rollback plan.

Phase 3: Network cutover planning

Network cutover is the highest-risk moment in a migration. Running both facilities in parallel (old and new) during the migration window gives you a fallback if the new facility has issues. This requires overlap rights in your old contract—negotiate them before you need them.

• Order cross-connects and carrier circuits at the destination facility after confirming timing with the provider before cutover
• Plan IP addressing: will you re-IP, use BGP to advertise existing ranges from the new facility, or bridge with a temporary VPN?
• Coordinate DNS TTL reductions 24–48 hours before cutover for faster propagation
• Plan carrier cutover: can you run both facilities in parallel during the migration window?
• Verify cloud on-ramp activation and cross-connect testing at the destination before moving production

Phase 4: Testing and validation

Test each workload immediately after cutover, before declaring the migration complete. Have a pre-defined validation checklist per workload class, and a rollback decision point if validation fails.

• Functional testing: every application and service passes health checks at the new facility
• Performance testing: latency and throughput meet or exceed the origin facility
• Network testing: all cross-connects, cloud on-ramps, and carrier connections are active and performing
• Security testing: firewall rules, access controls, and monitoring are correctly configured
• Failover testing: verify DR and backup systems work from the new facility

Phase 5: Decommission at the origin facility

After all workloads are migrated and validated, decommission the origin facility. Follow the exit terms in your old contract: de-install obligations, hardware removal, cable cleanup, and any restoration requirements. Do not cancel the old contract until all workloads are validated at the new facility and the migration overlap period has safely passed.

If the move is forced by a closure or a high renewal, see how to exit a closing or overpriced data center.

Contract terms that protect your migration

Negotiate migration overlap rights at your origin facility before you sign the new contract. Without overlap, you must cut over in a single window with no fallback—or pay month-to-month rates at the old facility while you migrate.

• Migration overlap period at the old facility (30–90 days to run both)
• Cross-connect and circuit ordering lead times at the new facility
• Early termination penalty at the old facility (ideally capped)
• Exit and de-install obligations documented in both contracts
• SLA enforcement at the new facility from day one of each cutover wave

Frequently Asked Questions