Systems management planning requires an arsenal of tools to provision, automate, monitor and control a modern enterprise data center. As data centers have grown and virtualization has abstracted workloads from the underlying hardware, the need for accurate and reliable software tools has become acute. IT administrators juggle multiple tools in order to gather a complete picture of the facility and its business-readiness.
The tools available for systems management now provide an avalanche of data, allowing administrators to delve into the status, setup and other nuances of any server. But interpreting the proliferation of systems data and making sound technical and business decisions is increasingly difficult, because busy administrators often overlook the role of time in systems management. Let's review the meaning of common systems management data in the context of time.
Tactical time horizons
A large portion of
Short-term tactical management data also frequently includes performance assessments of the business' various workloads. For example, reporting on the amount of CPU cycles, memory, I/O and storage actually utilized -- perhaps versus the amount of resources provisioned -- can potentially reveal a distressed workload causing a poor user experience. Technicians can use such insights to troubleshoot problems, streamline resource allocation and ease network congestion.
Another example may include system inventory and license management information, which provide IT professionals with a snapshot of hardware configurations, installed software versions and license deployment. These types of data are often captured periodically in preparation for an upgrade, to ensure the data center meets the minimum requirements for new product or patch deployments.
The key here is urgency; IT staff must respond to tactical management events promptly.
Strategic time horizons
Systems management tools can also produce a wealth of strategic data that should be used to make long-term IT and business decisions. Consider tracking workload performance over time. This offers more than just a momentary snapshot of each workload. Instead, IT administrators can review workload performance from a perspective of weeks -- even months -- to reveal cyclical trends that may escape tactical review. For example, a workload may demonstrate poor performance on the last day of the month, but only a long-term performance review might reveal similar poor performance on the last day of every month -- allowing IT to better grasp the cyclical computing needs of the company's payroll software and develop a more appropriate strategy for dealing with cyclical issues.
But analyzing workload data over a strategic time period also allows a business to perform workload capacity planning. For example, a single isolated spike in CPU usage doesn't warrant assigning more CPU cycles to a virtual machine, but when evaluating resource use over time, IT administrators may see a gradual escalation of resource use -- perhaps related to a growing number of users. This may lead to identifying a trend -- such as a dramatic increase in CPU usage over the next six months -- allowing the business to budget and deploy the additional resources needed to address the issue before it becomes a tactical problem.
This type of systems management data is all about planning, leading the data center and the business into the future.
There is no doubt that systems management tools -- and the data those tools provide -- are critical for successful data center management. But administrators must remember that different types of management data must be viewed in the proper time context for the insights to be most valuable to the business. It's all about urgency and relevance, so make short-term decisions based on tactical management data and make long-term decisions using strategic data.
This was first published in January 2013