Data center transitions are a complex undertaking usually requiring intense planning, a large budget, new infrastructure, and the ability to work across multiple departments and vendors. Most data centers are used for local resiliency or disaster recovery capabilities in support of production operations. Based on my recent project experiences, organizations should consider the following when conducting analysis and building out new data center capabilities:
- Business resources to accommodate technology. Often the biggest mistake is not focusing on the business resources to accommodate the technology. What good does it do to have the technical capabilities in place if the resources are not capable of performing the business operations? This can often lead to a difficult decision in whether or not to have redundant operations or at least a staff that has been cross trained in a different geographical region in order to adequately operate business processes. Be prepared to engage the business upfront in the process, considering the impacts such a move will have on their processes.
- Future performance and loads. Technology quickly becomes outdated—no surprise, right? To prevent your expensive technology upgrades from quickly becoming unsupportable, significant focus on future growth and performance needs to be applied when considering a data center transition. A good rule of thumb is to prepare for three years of expected growth.
- Customer-Facing Applications. Do your applications have external client-facing components? Often it becomes difficult to coordinate with external parties to ensure newly located applications can communicate with external applications and be tested appropriately. Make sure you’ve allowed sufficient time for thorough testing with vendors, and communicate as early and frequently as possible with them to align with project timelines at both the client and vendor.
- Retiering (Support Model) along with a potential retirement. This is often the case where applications in a different data center follow different support models. You should consider the remaining life of the application when planning the level of support it will need moving forward. Also, one should consider the upstream and downstream dependencies to the other higher-level supported applications. If my Tier 0 application is dependent on a file from my other Tier 4 application, then my Tier 4 application must have a workaround or must also be supported at similar levels to ensure the support can meet the demands of the business process in its entirety.
- Build and Deploy vs Copy Application. Does my application require me to perform a new build and deployment of the application and its components or can we simply copy over the server information for each component using jump servers and firewalls? Note that if the applications fall within the build and deploy category, there will be significantly more time needed to execute and perform transition of the application.
My goal in sharing these considerations is that you can run your data center transition in a more cost effective manner with minimal downtime by proactively addressing these factors in your analysis, planning, and execution.
I’d like to hear any feedback you have about these considerations and other thoughts you may have from your experiences in running similar projects.