Data center migrations aren’t something most people do every day. They’re typically a once-in-a-career event — twice if you’re lucky (or unlucky, depending on how you look at it). No matter which camp you’re in, moving networks, servers, data and applications from one location to another tends to elicit a string of four-letter words.
Slow. Pain. Ouch. Nope. (Not the words you were thinking?)
This is for good reason.
In helping hundreds of companies migrate everything from single applications to full data centers, we’ve identified seven common mistakes people make during data center migrations, and more importantly, how to avoid them.
A data center migration process can be broken down into seven steps, each with its own potential for mishaps: discovery, planning, development, validation, migration, management, and scale.
Mistake #1 (Discovery Phase): Lacking A Complete Infrastructure Assessment
The most common mistake made during discovery is not doing a complete infrastructure assessment. This is a rack-by-rack and U-by-U documentation of each device and its associated applications.
This assessment should note all things physical and virtual, network devices, network topography, etc. Don’t take shortcuts here because there’s no such thing as too much information for a migration.
Pro Tip: Include operational and technical interdependencies in your assessment. For example, app, web, and database servers on a related application must migrate as a bundle.
Mistake #2 (Planning Phase): Unclear Leadership
The most common mistake made during the planning phase is failing to establish clear leadership. This means identifying someone who is responsible for communicating clearly and definitively across all teams at all stages of the migration process.
A single department leading the way will, by default, look out for their best interest. The project leader must be an impartial party who understands and accurately reflects each group’s objectives and success criteria. This person must also have the authority to demand execution and the communication skills to keep everyone on the same page.
Mistake #3 (Development Phase): Not Recognizing Dependencies
The most common mistake made during development is upgrading parts of the infrastructure stack without recognizing the resulting dependencies. Of course, there’s nothing wrong with upgrading outdated components during a migration. New network equipment, for instance, is easily implemented, as are transitions from physical to virtual. However, when a migration has fractional upgrades, it’s easy to overlook the trickle-down impact of these changes.
Pro Tip: Be sure that any upgrades and their dependencies are noted for specific testing during the validation phase.
Mistake #4 (Validation Phase): Skipping Business Validation
The most common mistake made during the validation phase is skipping business validation.
During validation, IT, security, and network-engineering teams are typically heads down, hammering through checklists—and the business is ignored.
Changes inevitably happen due to unforeseen requirements. Application upgrades, for instance, have a habit of suddenly coinciding with a migration. Be sure the business understands all of the changes that can directly impact day-to-day operations. Investing 30 minutes here is going to save you days as you enter the next phases.
Mistake #5 (Migration Phase): Underestimating The Migration Timeline
A common mistake made during the actual migration is failing to set realistic time expectations. Production migrations are inherently slower than test migrations, as they require more care and attention to detail.
It’s always smart to pad your move time.
Pro Tip: Use the test migration to assess actual migration times. You’ll be surprised how long some applications take. Don’t just expect delays—plan for them.
Mistake #6 (Management Phase): Setting It And (Actually) Forgetting It
The most common mistake made during the management of your new infrastructure is adopting the “set it and forget it” mentality. Everyone is so excited to be done, they immediately wash their hands of the migration. You want to be sure you have both technical and business hands, eyes, and ears on everything while it settles.
Pro Tip: After completing the migration, plan on spending at least 48 hours on proactive monitoring and support.
Mistake #7 (Scale Phase): Thinking You’re Done
The most common mistake made during the scale phase is losing momentum now that you’re done.
In the words of Yogi Berra, “If you don’t know where you are going, you’ll end up someplace else.” These are words to live by when planning beyond your migration. Set an annual plan, maintain quarterly reviews, and develop a process for ad-hoc infrastructure requirements.
You’ve just invested significant time, energy, and money into executing a difficult (though critical) process. Don’t lose the energy or attention to detail now that it’s over.
This is not the only model for data center migrations, and these certainly aren’t the only mistakes people make. The important thing here is to constantly update processes as your technology, operating requirements, and experiences change.