5 Helpful Tips for Reducing Your Data Center Budget Without Harming Operations
It doesn’t take a lesson from the head of accounting to figure out that data center costs are on the rise. Technological developments mean new investments. Highly competitive markets mean boosting performance, productivity, and quality — all of which lead to more expenses. Is it time to trim the fat in your data center? Here’s how to do it without hurting operations one bit.
1. Lower Energy Costs
Since energy costs are one of the biggest cost-suckers in the data center, it pays to start cutting here first.
Energy costs are a significant factor in running a data center, primarily the energy it takes to cool the system components. There are some innovative ways to cut energy costs, though you might raise your eyebrows at a couple:
- Consider alternative cooling sources, such as cooling the data center from the free air outdoors. Or, try geothermal cooling or cooling from a local water source (such as a large lake or ocean). Paying for engineers to determine how to do it and running the necessary ductwork is a considerable cost up front, but can save significant amounts of money over the long term.
- Bump up the temperature from 68-degrees to 80-degrees. Okay, don’t quit reading before you hear it out. Traditional wisdom attributes equipment running long-term at higher temperatures to a shorter lifespan, which is true. However, replacing equipment a little earlier is less costly over time than cooling the equipment over the length of its life. Cooling the data center to 80 instead of 68 saves a lot of money over a long period of time, which offers greater savings than more frequent equipment replacement.
- Switch highly-active, read-only data sets to SSD. Costs of SSDs are coming down, they use less power, and generate less heat. So long as you aren’t using it for heavy read-write tasks, the SSD will last as long or probably much longer than HDD, saving money the entire time.
- Shut down unused servers. The commonly-held belief that powering servers on an off lowers the life expectancy is a myth. You can speed up the time it takes to power the server back on when demand is high by turning off unnecessary diagnostic checks. Users won’t even realize they’re waiting for a server to come on because the system will just slightly slow everyone down for a moment while the server powers up.
2. Lower Hardware Costs
Sure, the guys in HR are great. But do they really need 8 printers?
Can you rationalize the need for every piece of hardware you’ve got? Likely not. It’s probably time to retire a bit and eliminate the need for maintaining it, operating it, cooling it, and replacing it. This also makes asset management a lot easier.
3. Lower Software Costs
You can drastically lower software licensing costs by refining how and where you store and process data and workloads. On mainframe, you can offload resource intensive processing like Copy, SMS Compression and Sort processing to zIIP processers. Doing this allows you to pay less for CPU cycles while delivering better performance levels to their users. You can save more money by moving mainframe data from the expensive IMS databases into DB2 (which also addresses the challenge of dwindling IMS skill sets). Last, but certainly not least, you can target expensive storage and workloads from mainframes or data warehouses and offload them to Hadoop and Spark, saving more money and opening up the data to advanced analytics.
4. Lower Facilities Costs
Consolidation of data centers lowers costs in multiple ways. First, it eliminates redundancies of hardware and personnel. It simplifies initiatives like data integration, and significantly simplifies disaster recovery plans. It’s cheaper, easier, and more efficient to manage a single data center, even if it’s large, than to manage multiple data centers.
5. Lower Labor Costs
This is everyone’s least favorite way to save — but it deserves a mention. If you eliminate redundancies in hardware, software, and facilities, you’ll also need to eliminate redundancies within the labor force. Since this is the data center’s greatest expense, reduction can offer huge savings. It’s not always necessary to lay off employees — it can often be accomplished through attrition. Just don’t hire replacements when workers move, quit, retire, or transfer out. Of course, if there’s dead wood lying around, a good culling is probably in order.