Five Rules for Moving to the Cloud

Add Your Comments

According to IDC over one-third of all IT projects fail. Increasingly as companies move to the cloud, failure is following them.  As readers of this blog know, I am an ardent fan of the cloud. As I have said often, the cloud is for everyone, but not for everything. I want to discuss some of the hallmarks of cloud projects that are at risk.

1.  Your problems may follow you – You need to fix any application issues as part of or prior to the migration. In my current role at Cbeyond I see a number of clients that think moving an application to the cloud will “fix it.” The reality is, if an application has issues it is commonly the application, not the infrastructure. Unless you fix the application you are just migrating a problem to the cloud.

2.  Cloud does not equal 100 percent uptime – Often providers charge extra for geo-redundancy, take Amazon Web Services (AWS) for example. Unless you architect your application to be located in multiple regions your application may be susceptible to a failure at a single data center.

3.  It still takes smart people to make it work right – Unless you are using a SaaS or PaaS solution, there is significant configuration and management that you will be responsible for. It often takes a very skilled IT team to help run your applications in the cloud and not all cloud providers offer this type of service.

4.  Not all providers are created equal, pick the one that best meets your needs – Some providers assume you will be performing your own patches and backups, others do it for you. Additionally, some cloud organizations limit support to issues with the hardware and OS, while others provide application-level support. You need to fully understand your needs in order to select the best provider to meet them.

5.  Copying your current specs to move to the cloud can lead to failure – Often clients approach my team with “this is what I have today, please give me the same specs in the cloud.”  While this can work, without an understanding of how the application is running today, the type of connectivity the client is using, and expected growth rate, it can be a very dangerous way to scope an application.

Cloud projects can fail for a number of additional reasons, the problems I have listed above are a few of the more common examples. The best way to ensure success for a new cloud implementation is proper planning, along with developing a strong understanding of your applications and cloud vendor(s), to make sure they are a good fit for one another.

Add Your Comments

  • (will not be published)