Everything needs to be revamped at some point, even data centers. In fact, data centers become obsolete faster because the state of technology advances so quickly. Not only do servers, data storage systems, power supplies and other equipment wear out, but by the time they are ready to be replaced, you could be generations behind the latest innovations.
Deciding when it is the right time to upgrade your data center design is a matter of balancing the replacement of aging systems before they fail and the need to keep pace with the latest computing trends.
According to a recent survey of government IT professionals, data center operators need a more effective decision-making process for data center migration. Seventy-four percent of respondents say that government agencies hold on to outmoded data center systems longer than they should, and 62 percent say they need more effective data center migration strategies. The three biggest reasons IT managers choose not to upgrade their data center design are cost (51 percent), the need to control critical processes (46 percent) and the time required to migrate existing systems (41 percent). However, every data center will need to be redesigned at some point. The question is, when is the right time?
Assessing Data Center Lifespan
There isn’t a specifically calculated or scheduled time to upgrade data center design, but you want to have a migration strategy in place before you need it. The average lifespan of a data center is 10-15 years, but specific elements within that data center will need to be upgraded much sooner. IT consultants AKF Partners offers a breakdown of the average useful life of various data center infrastructure components:
- UPS batteries – 4 years
- Diesel generator – 30 years
- Programmable logic controlled switchgear – 15 years
- Computer room air conditioning/handler (CRAC/CRAH) – 12 years
- Galvanized steel cooling tower – 10 years
- Electrical distribution board – 25 years
- Water piping for chillers – 30 years
However, this doesn’t include the computing hardware. Servers and computer systems wear out based on use; the more data they have to process, the faster they wear out. The rule of thumb is that servers should be replaced every five years, although hard drives and internal components can wear out faster. Solid-state drive systems (SSDs), for example, degrade with every write operation—the more use the drive gets, the faster it degrades.
Clearly, the question of when to upgrade your data center design is a factor of diminishing returns on component replacements, as well as the need to keep pace with new technologies. Depending on your organization, any number of factors can trigger the need for a new data center design.
When You Know It’s Time for a Data Center Upgrade
Demand for more data is what is driving the need for new, more efficient data center designs. As companies consume more information, the need for data storage capacity increases exponentially. Even with support from cloud data storage, most organizations need more localized data storage capacity to accommodate faster data processing and analytics, which can be accomplished by eliminating data transmission latency.
The increase in demand for more data storage directly correlates with company growth. The faster companies grow, the more data they need to ingest and the more data center capacity they need. The data center market is expected to grow by 12- 14 percent in the next few years, but half of that data traffic is expected to be handled by a handful of the biggest data facilities. As companies assess their own growing data needs, they need to determine how much additional storage capacity and physical space they need on site. More companies are meeting their data expansion needs by leasing more cloud space for non-critical storage, although they continue to maintain local data stores for business-critical data processing.
According to a recent Forrester report, even though companies are adding more data storage to accommodate bigger workloads, “they retain aging hardware for four years on average, which is longer than ideal.” Continuing to use aging servers limits business agility. As data demand increases, you have to assess current hardware specifications and performance parameters so you can stay ahead of capacity limits
Slowing performance is usually a sign that it’s time for a data center upgrade. The effects of data gravity require that more data be stored locally to reduce latency and processing lag time. As more real-time applications driven by machine learning, 5G and the Internet of Things (IoT) become part of the corporate computing fabric, increased data capacity will help companies deal with the need to move data processing closer to users.
How to Allow for More Data Center Space
Ideally, your data center expansion should have been included as part of the original data center design and planning. One of the biggest mistakes data center architects make is failing to match design criteria to performance characteristics. They can oversize design, providing more capacity than is needed with associated costs for underutilized space. Ideally, the original design was right-sized for maximum operating efficiency, leaving a path to system upgrades.
Flexibility and scalability need to be central to the design. That includes electrical systems, so you can add input and output distribution systems to handle changing computing needs and being able to increase cooling capacity without disruption.
Projections for company growth should also have been included as part of the initial design criteria, offering a path for expanded capacity. Of course, growth can be difficult to predict. Using a modular system design can smooth the way by making it easier to expand the current infrastructure; adding capacity as needed without placing an added strain in HVAC and power—because the modules are self-contained systems. More data center engineers are using modular to simplify data center design and expansion, because modular designs can be engineered to fit the space available, and they provide a standardized, repeatable design for expansion.
There isn't a specifically scheduled time to expand your data center. However, when the time comes, symptoms such as degraded performance and overextended data storage will make it clear it’s time to upgrade. You can be prepared by utilizing modular designs as part of your data center build-out. Modular systems can be fabricated and installed faster than individual components, and they deliver predictable performance to preset specifications.
If you want to learn more about modular systems and how they are impacting the future of data center design, be sure to read our latest e-book, The Complete Guide to Modular Data Center Solutions.