Data centers today need to be built to enable growth, which means they need to take advantage of virtualization technology and cloud computing models, and they should scale for globalization to capitalize on growth opportunities. They need to be able to ensure the availability of the applications and networks they support, and perhaps more importantly, protect the data those applications create. Finally, they need to operate efficiently while facing the myriad of new technologies replacing legacy architectures. In short, they need to factor in the following requirements in their march toward modernization:
- Data integrity and availability
Capitalizing on these requirements and their associated benefits is no easy feat. IT often operates in a silo from other business departments, which challenges today’s application service models. Replacing legacy architectures can be risky, and downtime is becoming more and more costly. Therefore, it’s important to fully understand the costs of modernization and downtime in advance.
One of the areas in which modernization often hit roadblocks is with the increased complexities found in today’s data centers. Four big lessons we’ve learned in our data center modernization efforts include:
- A company has to periodically scrutinize its infrastructure and rid itself of unnecessary and costly platforms. If this step is not performed, the value proposition of the product or variant will eventually shrink. At the same time, the costs necessary to manage the complexity, which are also difficult to reduce, increase at an above average rate.
- It’s best to focus on major applications as a starting point with this process and to involve key stakeholders on both technology and business sides. Organizations will get a better success rate when this is accomplished as a business initiative where key business and technology leaders create an enterprise-wide vision.
- Executive support and buy-in from technology and business sides are key for a successful application mapping assessment. Selecting the right team players for the assessment plays a significant role in making it a success. In our experience, the application business lead needs to have an entrepreneurial mindset that can drive this initiative in achieving its goal of increasing the IT infrastructure agility.
- Don’t underestimate tribal knowledge. Tribal knowledge is the highly-specialized, unwritten, and unshared knowledge that employees collect throughout their employment. It can be an inherent risk to IT departments because it’s difficult, if not impossible, to discover using automated discovery tools or to capture manually through formal processes. The existence of tribal knowledge is especially apparent when implementing new data center technologies. As data centers become more complex, it’s not at all unusual for server versions to differ from instance to instance or for certain settings to be changed to improve performance. At any given point throughout your data center, engineers have probably made slight modifications that they never documented.
Capturing this tribal knowledge can be extremely difficult. First of all, it’s usually spread throughout an organization and the owners of this knowledge aren’t always available when the information is needed. That’s why people with tribal knowledge get called into the office on weekends and at odd hours—because someone is encountering unexpected difficulties while updating or installing new technologies.
Ultimately, the risks associated with these lessons add to the overall fear that data center modernization will be fraught with excessive downtime and lead to projects that either never get off the ground or are never fully realized. But it doesn’t have to be that way.