Your data centre is the backbone of your IT Infrastructure, but with any line of your business there will come a time when you need to migrate to a new one.
You could be undergoing a merger or acquisition, your on-prem facilities could just be too old or you could be looking at new facilities to save money.
Undertaking a data centre migration project can produce enormous benefits to an enterprise, yet the process can be risky and expensive.
Business Challenge: Data Centre Migration to Save Money
It Naturally advised a customer that a network transformation, migrating their core network services from one data centre to another would save them around £500k per year.
The data centre migration involved moving MPLS and VPN terminators, firewalls, DNS and other services.
The successful network transformation project culminated in a single big bang network change that had to be delivered within a critical outage window.
Here is our check list to ensure success of a critical cutover
- Early engagement of business stakeholders: If working with an IT Managed Service, as the customer, you should be involved throughout the project from start to finish. Your involvement is crucial from the start to begin planning. the outage window which will be the least impactful to the business.
- Check and validate everything: Check the work and assumptions of all parties including third party SMEs. Errors do occur and often third parties follow process only rather than think through the eventual outcomes.
- Firewall rules: Check, check and check again.
- Utilise a detailed runbook: In the weeks running up to a change, IT Naturally use a templated runbook to capture step-by-step the technical implementation activities, durations and responsible task owners. It took several meetings and revisions to ensure this is correct.
- Robust cutover entry and exit criteria: Before the implementation is started and approval given to proceed, a detailed checklist should be used to ensure all technical, service, communication, stakeholder and resource preparations had been made. It’s a bit like getting prepared for takeoff and then there’s the landing. Robust exit criteria should also be defined and approved in advance by your company stakeholders. This agreement allows the change to be completed and that the required outcomes will be delivered and that the service will still be supported.
- Stakeholder and sponsor support during the change: You should expect that your key business stakeholders and sponsors are present to sign off the entry criteria. Their time should be pre-booked for checkpoint calls during the change to enable rapid decision making and approval if required if anything needs to deviate from the already agreed runbook.
- Agree rollback criteria, triggers and timing before change starts: Your runbook should clearly indicate and allow time for rollback, to ensure business operations will be enabled within the change window.. Plan and pre-agree with your stakeholders for an unsuccessful migration detailing why and how rollback would happen.
- Ensure everyone required is confirmed: People from all parties should be included from the planning stage and clearly communicated in the runbook. Confirmation of the availability of stakeholders and engineering resources for secondary system should be on call.
- A team with legacy knowledge: Where possible ensure the team on hand have as much background knowledge as possible on the history of the systems being changed or potentially impacted.
- Your team to test resources: It’s invaluable that your people are on hand to functionally check all impacted systems following the network migration and confirm exit criteria had been met.
- Log and robustly prioritise issue resolution during the change: With any large project there will be a few problems. In our latest data centre migration, IT Naturally hit a few snags relating to firewall changes that had to be resolved during the outage window. We found clearly capturing these and focusing all parties on the resolution of each snag one at a time was instrumental in resolving all issues within the outage window.
- Communication is key: Technical teams using a telephony bridge and messaging service to discuss issues, share technical information and screen shots is a fantastic way to track progress. A separate communication messaging channel is great for your business to keep you team updated on progress. Regular meetings should be scheduled with your stakeholders should their support and governance be required.
- Understand all potentially impacted systems: Before starting the implementation ensure all potentially impacted systems are understood and listed out. Ensure technical documentation for each is available if required to troubleshoot.
- Testing: Successfully testing each of the listed potentially impacted systems must be part of the acceptance criteria.
- Two sets of eyes are better than one: For all technical changes, have a primary resource making the changes with a secondary knowledgeable person watching and checking.
These are the 15 steps that IT Naturally follow for all our data centre migrations. It is a complex process so do ask the experts to help you complete yours.