3 Keys to Effective Data Center Capacity Planning
There is a well-earned sense of satisfaction when your IT department finally has your data center set up and performing the way it’s meant to: efficiently and securely managing your software applications, telecommunications, and data storage. However, this is not the time to rest on your laurels. Instead, you need to formulate a data center infrastructure management (DCIM) plan, a major component of which includes data center capacity, to account for current and future needs.
Determining current and upcoming hardware capabilities falls into two categories: inventory and performance. Knowing exactly what you have will allow you to understand exactly what your data center can (and can’t) do, as well as assess its future capabilities.
On the inventory side, you need to document each server’s name, make and model, operating system (OS), speeds, memory and disk space, network interface cards (NICs) and speeds, and total number of central processing units (CPUs). On the performance side, document each CPU queue, percentage per CPU used, percentage of memory used, memory paging, disk inputs/outputs (I/Os), and actual network speed.
While the data center you have now may work fine for your current needs, what happens if (or when) you have to scale up? Costs to run a traditional data center typically range from $10M to $25M per year, and you’ll likely need to investigate—regardless of the system you are using now—the possibilities offered by the four types of data centers:
- In-house facilities owned and operated by the company using them
- Co-location systems where servers are housed together for use by multiple clients with their own private computer cloud storage systems
- Scientific computing systems run by national laboratories
- Public cloud providers such as Google and Amazon
No matter which of the options above works best for you now, or will be your optimal choice in the future, the decisions you make should consider the following criteria:
- Fault tolerance (fire suppression, air conditioning redundancy, dual electrical sources)
- Security/safety (biometric hand scanners, defined floor access procedures, locked cages and cabinets)
- Capabilities/scalability (total cooling capacity, minimum 18” raised floor, number of cabinets)
- Core service standards (on-site service personnel, maintenance schedule, alarm system)
- Location risk profile (flooding probability, aircraft flight plans, distance from railroad/highway, seismic profile)
Modern data center capacity planning employs a needs-based approach utilizing just-in-time upgrades and expansion so expenses are incurred only when needed and capacity meets current needs with the opportunity for future growth. And, yes, it can be hard to calculate and project these overlapping variables on your own. So when creating your DCIM plan, don’t reinvent the wheel from scratch. Instead, contact Silverback Data Center Solutions about their hosted DCIM packages and managed data center solutions.
Comments are closed