As you increase use of public cloud resources, how do you make sure that you're winding down and consolidating the data center to optimize performance and cost? Data centers are expensive to maintain. It's not uncommon for an organization with three data centers to spend millions annually on these facilities.
With the proliferation of cloud computing platforms like AWS, Azure, Google Cloud Platform (GCP), etc., many organizations are moving workloads out of data centers to reduce their facility footprint while achieving greater agility in how they deploy applications.
So why is data center consolidation harder than most people think it will be?
Infrastructure Lifecycle Challenges
The data center infrastructure lifecycles are often out of sync with the cloud migration timeline. Many organizations treat their data centers as a long-term investment, deploying new technologies over time and retiring older systems only when they reach the end of their useful life.
A problem with this approach is that years of amortizing hardware and facilities costs are unlikely to be aligned with your planned movement to the cloud. This results in having vacant on-premises infrastructure that is going to waste, or having application consolidation dictated by hardware lifecycle changes rather than by the choice of the ops and application teams.
Purchasing, leasing, and renewal of hardware is designed for financial advantage (e.g. depreciation, chargeback alignment, hardware vendor discounting, burn down of large contracts). This means that your application architecture may be as much designed by your CFO as it is your development and ops teams.
Application Availability and Business Continuity
Applications are exposed to hardware failures and data center outages no matter where they run. But on-premises infrastructure and application resiliency is designed by your teams, not by a simple API call to a new region like you have with the public cloud.
Consolidating down data centers must take into consideration the availability and business needs of the application architectures. This could lead to a longer timeline as you work on application refactoring and re-platforming which impacts how you make consolidation decisions.
The data center ops team is used to running the data center in a certain way. They have responsibility for capacity planning, performance analysis, security, patch management, and day-to-day operations. Consolidating data center infrastructure may require changing a lot of these procedures and rethinking ops processes. Change is harder than most people think it will be.
Application Development and Lifecycle Challenges
The application teams are often used to deploying new applications on-premises and not in the public cloud. The move to the cloud often means a change in how they develop, deploy, and run their applications.
For an organization to successfully consolidate data centers, the application development and ops teams need to be on board with moving workloads to new design and operational patterns in the public cloud and balance that against the constraints of winding down on-premises resources. Decisions must be made in concert with data center designers, operators, and facilities operations teams that align with the direction of the business, especially in light of significant workforce changes that have transpired in the past 24 months.
Data Center Consolidation is a Complex Challenge
The concept of consolidating data center resources seems simple, but presents a complex challenge that requires input from all aspects of the business, including finance and facilities.
It's important to factor in application architecture changes as well as operational needs when making decisions about data center consolidation projects. It also means thinking long-term so you can have a vision for how your on-premises infrastructure will be bought, operated, and removed, to meet evolving and fluctuating requirements.
There is no one-size-fits-all answer. It's important to take the time to properly plan your data center consolidation project so you can ensure it meets the needs of your organization and your applications.