Lift-and-shift cloud migrations are dying

According to Pluralsight’s recent State of Cloud report, 75% of IT organizations are building net-new applications and innovations in the cloud. That means 25% of their applications undergo lift-and-shift migration.

There’s a debate about the disconnect when it’s time to execute cloud migrations. Lifting and shifting application workloads is known to limit the benefits of being on a cloud platform in the first place. The shifted applications do not take advantage of cloud-born features such as serverless or cloud-native features such as Kubernetes and containers.

Lift and shift was once the most popular way to move applications and data to the cloud and it remains popular with many enterprises. The idea is to basically replicate the platform on a public cloud provider. Is there a better way today? What benefits are we missing by using a lift-and-shift approach?

Enterprises need to modernize applications to optimize them for the cloud platforms they reside on. This was viewed as costly and unproductive by most enterprises that valued speed over efficiency. Indeed, it was the norm during the pandemic.

Even enterprises that initially did more refactoring during migration (optimizing them for the target cloud platforms) fell back to lift and shift to speed migration to the cloud. At the time, enterprises considered systems that remained on premises at higher risk since many pandemic shutdowns also limited access to traditional data centers. This gave IT a license to move faster, and that meant skipping modernization steps such as application refactoring for the target cloud platforms.

It seems we’re now paying the price. If you look at the recent surveys, as I’ve covered here, cloud costs are far higher than most enterprises expected. At this point, boards and executive teams may be shutting down cloud computing growth, at least until they can figure out what’s wrong.

Today, the thinking in most enterprises is that we need to slow down to go faster. This means investing in refactoring applications to reap cloud-native benefits. Refactoring also produces applications that are less expensive to operate.

Most cloud sticker shock I see these days is due to a lack of cloud cost monitoring and optimization (finops), and the fact that most lifted-and-shifted applications run like dump trucks when they should handle like a new Tesla. Of course, the larger issue is that the business is impacted, for many to a point where core business failures may be traced to the enterprise’s inability to leverage cloud computing for what it should be—a true force multiplier for the business.

The bottom line is that most enterprises leave money and business opportunities on the table when they lift and shift applications. What’s worse, they don’t even know they’re doing it. They’re confused when the business goes south and the move to cloud computing only made things worse.

Should lift and shift no longer be an option? Of course not. All options should be on the table. Lift and shift is fine for applications that won’t benefit from cloud-native features. However, it can no longer be the go-to solution for enterprises looking to quickly move applications to public clouds.