It wasn’t too long ago that the idea of migrating an infrastructure to the cloud was seen as a gamble. Cloud technology was still in its infancy as recently as 2009, and companies that made the jump from on-premise data centers to the cloud were either pioneers of brand new tech or suckers buying into the hype, depending on who you talked to.
Nowadays, cloud-based applications are the norm for startups and newer SMBs. It’s a low-cost, flexible, and scalable alternative to a personal data center, a solution that lets companies adapt to changing needs on a dime without worrying about the upkeep and maintenance of on-premise hardware.
But just because it’s extremely popular doesn’t make cloud migration any less overwhelming. For older companies, ones that have relied on on-premise servers for years, the move can feel like a total upheaval. Maximizing your migration to the Cloud can mean re-thinking the way your business operates and organizes.
Crawl Before You Run
One of the most difficult aspects of cloud migration—especially for older, companies—is that their current legacy tech is part of a monolithic system. Its elements are tightly interwoven, which makes it hard to swap out components without disrupting the entire machine.
That’s the difference between cloud applications and monolithic, legacy ones. The Lego cube represents cloud services like AWS, and the concrete one is the legacy tech.
‘Lift-and-shift’ describes the simple transition from data center to cloud, with little to no changes in how the application works or runs. It may not be optimized for the cloud, or leveraging the cloud to its maximum potential, but that’s fine for now. Lift-and-shift is quicker, easier, and ideal for simpler applications (where there aren’t too many major differences in how it runs between the cloud and on-premise).
‘Fix-and-shift’, on the other hand, takes the time to optimize and reorganize the application (and its workflows) before it’s deployed. This doesn’t just mean re-developing either—it can extend to re-structuring entire teams.
Which metrics you’ll actually be monitoring, and the fidelity of your simulation, will depend on the specific type of cloud application and its use case. But as a rule of thumb, the more you can match the final environment, the smoother your launch will be. Better to just tweak to the resources you need than to go-live with a rough estimate that turns out to be more wrong than right.