Workload Optimization Series: How to avoid cloud migration mistakes

It’s not about taking discrete steps, it’s more like a journey. Organizations that don’t view the process this way risk making mistakes, such as spending more than they had budgeted, or having their applications under-perform. The following are some of the biggest cloud migration optimization mistakes we see and how you can avoid them.

• Don’t look at cloud as just another infrastructure. Some organizations think they can move their workloads from a data center to the cloud and instantly reap benefits. When companies take this “lift-and-shift” approach, their costs can rise significantly. IT managers who run workloads in the cloud the same way they ran those workloads in an on-premises environment are going to over-provision and pay more than they should.For example, an on-premises workload is typically provisioned so it can run properly at the most demanding times. As a result, the infrastructure is over-architected most of the time. A retailer, for instance, will set up their infrastructure with an always-on approach so their workloads can run during the peak November/December holiday shopping period. In the cloud though, you don’t want, or need, to over-provision. It can get expensive. Instead, you should look at potentially scheduling workloads for different times.
• Recognize that not all data belongs in the public cloud. Many CEOs love the idea of a cloud-first strategy because it allows them to turn capital expenses into operating expenses. But the reality is some data can’t be moved into the public cloud. For example, data that needs to remain resident in one country due to privacy or regulatory requirements isn’t a good public cloud candidate. Another example would be data with strict latency requirements, where it might be best to have the data as close as possible to where the users are located. Make sure you understand your data and all its requirements before you decide to migrate it to the cloud.
• Not all applications should be moved to the cloud ‘as-is’. Some large, monolithic applications won’t operate optimally in the cloud. In many cases, it’s preferable to re-architect such applications using micro services and containers. Then you can run each portion in a smaller footprint and scale horizontally if you need to. This way, if you start to run low on resources, you can spin up new resources automatically and pay only for what you use at any particular time. This ability to scale on demand is one of the cloud’s greatest strengths, but it only works if your applications are architected to suit the cloud.

Migrating to the cloud takes technical knowledge, planning and preparation, but it can ultimately be very beneficial for businesses. If you’re not sure you have the skills available in-house for cloud migration optimization, a trusted Managed Service Provider like Aptum can be a great resource to fill in any knowledge gaps and provide additional support.

Read more in our Workload Optimization Series: Why different workloads require different infrastructures