A recent survey is telling us what most of us already know. Cloud migrations are complex, expensive, and often need to be more successful. Perhaps it’s time we stopped digging a migration hole and learned to do better.
Cloud migration is a significant effort in the cloud computing world, even with the advent of generative AI. Cloud migrations are a core effort for data transformation and moving applications to platforms that most enterprises believe are “future-proof.”
After 10-plus years of migrating applications and data to the cloud, you would have thought we could get it right by now. However, that is not the case.
- Network bandwidth issues impact employee connectivity, which is an obstacle in the new world where employees are scattered hither and yon, some without reliable internet connectivity. Adequate bandwidth is mandatory to take advantage of “the cloud.”
- Latency or slow application performance is thought to cause quality issues. This negative experience increases users’ frustration. I’m sure this is also related to bandwidth issues.
- Downtime and poor user experience are causing companies to incur losses in productivity. I suspect this is not due to cloud outages but other ignored issues, such as network outages and system components down for maintenance. Many people fail to account for this.
- Finally, there are limited IT resources to assist with migration challenges throughout each phase. This comes down to the need for more talent in the enterprise and some employees who no longer want to work on migrations.
The core of the issue
The reality is that we’ve not done migrations correctly, for the most part. The cloud providers sold the cloud as something that needed to be leveraged ASAP, so massive workloads and data sets were lifted and shifted to this new “miracle platform.”
Three things occurred:
First, it was more expensive than we thought. I use the unproven number of the cloud costing 2.5 times what enterprises believed it would cost to operate workloads and data sets in the cloud. This all blew up in 2022, when we also had the accommodation of workloads moved during the pandemic, many with unimproved applications and data sets.
Second, poorly designed, developed, and deployed applications moved from enterprise data centers to the cloud, where applications still need to be better designed, developed, and deployed. We’re paying more for them to run in the cloud since we’re paying for the existing inefficiencies.
The cloud is often portrayed as a platform running elegantly designed cloud-native and container-based systems. Most workloads running on the cloud are unmodernized, and although they function, they are underoptimized. In many instances, they are sold to leadership as future projects where the application modernization would occur. Most of this just turned into technical debt, which means kicking the can down the road while paying for the road.
Finally, enterprises aren’t learning from their mistakes. I’ve often been taken aback by the amount of lousy cloud reality that most enterprises accept. Although some have moved back to enterprise data centers, some are indeed funding application and data optimization. We’re still getting a C- in returning value to the business, our shared objective.
The fix
There is no magical tool. Leveraging cloud-native approaches for most applications will be counterproductive. Success will require great attention to planning and understanding where things are now, where they need to be, and steps 1 through 100 to get there. This is often too complex, scary, and expensive for enterprises to consider in 2024.
The hard work needs to be done now. I urge enterprises to focus on what needs to change and what needs to be done to make that change. It’s not hard to figure out, and I suspect most enterprises understand that this must get done.
The catalyst for fixing things now is the massive interest in generative AI. This game-changing technology will allow many enterprises to penetrate deeper into their markets. This won’t occur considering the current technical debt and the reality that you’ll either do everything you need to do or pay for the jump into the generative AI pool. Otherwise, don’t bother.
However, this means a commitment to change, which means first admitting you have a problem. Who will be first?