Don’t migrate your problems to the cloud

The cloud is typically a destination for systems needing to be modernized to take advantage of technologies such as AI, predictive analytics, or a hundred other cloud services. It’s typically cheaper, it can be allocated and changed in minutes, and the enterprises technology elites are spending most R&D dollars on the public cloud these days. Thus, your existing platforms are no longer getting the love.

Moving to the cloud is not a bad idea. However, the trouble comes when enterprises believe that digital enablement will somehow fix existing problems, such as a data mess, application issues, inadequate security, or frequent outages due to a lack of operational disciplines and tools.

These issues won’t fix themselves once applications and data migrate to the cloud. You’ll end up with less-than-optimal systems in the cloud, made worse by the fact that those working in IT will still be on the learning curves of using public clouds. In other words, you’ll be increasing risk and cost without improving applications and data. Things will likely get worse. 

I’m not a fan of migrating problems to the cloud. Indeed, I warn people of three consequences of migrating data and applications that are problematic.

Security and outages. If your data and applications were less than secure on-premises, migrating them to the cloud will likely not improve things.

If we try to rethink security in the “as is” state, perhaps leveraging identity access management or upgraded encryption, then it’s just a matter of finding analogs in the cloud. The alternative is rethinking and rejiggering security on cloud platforms where those not as experienced in the cloud are more likely to make mistakes. Same goes for ops.

Data, data everywhere. If you’re normal, chances are you don’t have single sources of truth for your enterprise data. You likely have a massive number of redundant databases, and worse, no one really understands where all the data is and what it means.

In migrating problem data to the cloud, chances are you won’t migrate all data. Part of the movement to cloud is figuring out the integration of the data from on-premises to the cloud, as well as how the data is leveraged differently by each application or user. You’ll just make things more complex, and the situation worse.

A culture that’s not ready for cloud. The number of failed migration projects that can be traced to culture issues is beginning to increase sharply. Many well-intentioned IT leaders focus on moving stuff quickly to the cloud (especially during the pandemic) but not on the humans who deal with the “as is” and the “to be” states of the systems, off and on public clouds.

Although it’s difficult to change a culture, it’s not impossible. However, it cannot be done during the months it takes to lift and shift data from on-premises to the public cloud. As a rule, retraining, hiring, and changing hearts and minds takes at least twice the time as migrating major systems and data to the cloud.

This is not that difficult to grasp. Basically, the rule is: Fix problems before you move or don’t move.