What are some concerns companies might have with a "lift and shift" cloud approach?
The process of “lift and shift” for some set of workloads or data environment is commonly defined as a project that simply seeks to “lift” the operations out of some environment and then “shift” it to another – for example, a move of certain workloads and tasks from on-premises to off-premises, or a move of data operations from one data center to another. The concerns that companies may have involve some of the nuts and bolts of how such a migration might work.
Some concerns with lift and shift relate to the idea that it’s not as simple as it sounds to just “cut and paste” system operations to a new system. Often, a lift and shift approach will start without sufficient documentation of requirements, or operational design. When these are required later, it can throw a monkey wrench into the project.
Other problems with lift and shift involve the differences between the two environments in question. A project might work fine in an on-premises or original legacy system, because that system has all of the right resources, but it might not work as well in a new location. An oversimplified analogy would be to a houseplant: there are any number of factors that might make a plant thrive less in one pot than in another.
Other concerns with lift and shift have to do with inherent problems in the original process: there’s often the idea that in a lift and shift, there’s no effort to fix problems before migration. That leaves those problems bedeviling the project in its new environment, and causes rampant chaos as stakeholders are trying to implement and scale the new system. VMware vExpert Eric Wright, writing for the Turbonomic blog, makes an analogy for this problem as well: the analogy of a moving company that simply “lifts and shifts” containers full of trash or debris, delivering junk components to a destination. The idea with IT lift and shift is that the results may involve useless jumbles of data, data that is mismatched to its handling systems, or data sets that outgrow their “habitat.”
All of these are potential problems with a “lift and shift” scenario, which can be contrasted with other scenarios called, perhaps, “fix and send,” where there are upfront efforts to design and perfect systems before a migration.