Data center design continues to be a hot topic, as technicians and designers seek to squeeze the maximum amount of power and storage into a finite space. Data centers come in various types to meet business needs, with a constant stream of improving hardware, new design philosophies and growing business to be considered.
While hyperscale data centers will always look to cram in the power for a wide range of customers, most enterprise data centers only need to service the needs of that business, and can use more predictable growth statistics and business plans as the basis for capacity planning.
The importance of data center capacity planning
Whether a new-build data center or an upgrade, data center capacity planning needs to be a key part of the design process. Capacity planning used to involve considered estimates about business and data volume growth, but has become more of a science that includes building a strategy to ensure overall computing resources meet future business requirements, including power, cooling, storage and management capabilities.
As your business grows, you must better understand the changes in its data center needs and plan accordingly. Given the cost and investment in data centers, getting the strategy for growth correct is of key importance, especially as clouds, edge computing and other trends impact how data centers are utilized.
As enterprises mix up their workloads, having a data center that can cope with the rebound in data from a failed cloud project, network outages and other problems are a sign of a progressive and efficient organization.
Best practices for data center capacity planning
When designing a new data center, rigorous planning is required both to get the right technology in place and ancillary power and cooling services for a cutting-edge facility that will support the business for years to come. But having a data center hit maximum capacity shortly after launch is a sign of poor planning in an environment where mergers and acquisitions activity can see businesses merge huge sets of data, or worker and data growth continues unabated.
With most data centers taking one to two years to build, the storage, processing and network hardware must exceed business needs and be upgradable to cope with future growth. Additionally, the data center needs to be easy to manage, making those future plans simple to cost and justify.
Fortunately, data center infrastructure management (DCIM) applications help take some of the strain when it comes to capacity planning. These provide a clear view of current operations and can highlight bottlenecks, spikes in usage, growth in data and network demands through dashboard views.
Using DCIM within a new or existing data center automates most of the measurement tasks, highlighting failures, errors and likely breakdowns early and improving uptime thanks to the smart lifetime monitoring features of most components.
Effectively employing DCIM can support current capacity utilization and planning for upgrades and boost data efficiency improvements.
7 steps to effective data center capacity planning
For those migrating to DCIM or using other methods for their capacity planning, follow these seven steps to deliver success.
- Create a baseline of current data center performance, if available. Take an inventory of components and consider which are already close to retirement.
- Establish existing and future staffing, data and usage growth patterns and explore the business strategy to see where further boosts in data demand are likely to come from.
- Agree on metrics, technology selection and budget across the IT department, leadership and other key roles. Choose the right technology partners where needed who can deliver on your business results and needs.
- Look at what can be upgraded in situ (larger SSDs or hard disks, processors and memory) or where the data center needs to be expanded or reconfigured to support new aisles. Alternatively, consider if a new data center or other solution is needed.
- Use DCIM to monitor the day-to-day and longer-term changes in usage, and align data staffing around greater automation.
- Revisit the data center capacity plan regularly to see if elements of it must be accelerated, can be delayed or need to be updated.
- Check that your growth figures remain accurate and within boundaries to ensure further decisions are based on solid information.
Outside of best practices, the IT team needs to monitor data center trends to consider adopting new storage technologies, supporting new networking paradigms and other shifts.
As data centers look to be more efficient, with additional power provided by solar, wind or other green methods, they can help an enterprise with their environmental credentials. Some already put their waste heat to better use, and as businesses become more reliant on them, a scalable and modular design may take priority over previous requirements to ensure the data center can continue to grow in step with the business.
Despite the hype about the cloud, many enterprises will still demand their own data centers for compliance, security and speed or productivity reasons. But those data centers will need to be almost as scalable as cloud services, as business users expect similar levels of flexibility and departments come up with new systems that create more data or need extensive resources to crunch it as part of big data or AI-based processes.