Companies across all industries are turning away from physical storage and computing power to cloud-based resources. This is good news and bad news for data centers. On one hand, more business for data centers — on the other, more demand and strain being put upon limited resources and hardware.
As demand for cloud-based computing power soars, it's not unusual for data center managers to find themselves stretched thin.
Luckily, there are practices that data center managers can implement to improve the efficiency of a data center and help data center workers better utilize the hardware they already have. Here are 4 ways to improve the utilization of a data center.
1. Keep Your Network Charted
Creating a network diagram that properly charts the topology of a data center — depicting the relationships between network switches, cables and servers — is a critical step in ensuring that data center's efficiency of space and resource usage.
Network topology diagrams can also reduce training costs or time, as they provide a way for new employees to quickly get up to speed on both the overall topology of a network — like three-layer, leaf-and-spine or FatTree — and how the different components in a particular data center fit together.
Network mapping can be time-consuming, and is often considered one of the most tedious parts of networking — but there are ways to make the process simpler and easier.
To be the most effective, a network diagram will need to be updated regularly — but the benefits outweigh the time and labor needed to create and update these topology diagrams. Keeping a network diagram up to date will make it easier for data center managers to keep track of equipment, switch to another topology if necessary and optimize existing resources.
2. Pull the Plug on Zombie Servers
Some studies of commercial data centers have found that more than 30 percent — nearly a full third — of servers in data center operations are zombie servers, servers that have shown no signs of useful computing activity for six or more months.
If left unchecked, zombie servers take up rack space, use energy and hold cables in place. These servers use energy and take up rack space while providing no value for the data center that houses them. As much as $30 billion in possible profit is wasted each year on maintaining these unused servers.
Regular review of a data center's assets can help managers ensure that all of their online servers are providing useful activity. When a data center finds a zombie server, they can pull the plug, then plan for how the server can be reintegrated.
In a similar vein, taking regular stock of non-server equipment — like cables and switches — can help managers ensure that they have a good idea of their data center's equipment stores. This can help prevent managers from purchasing new cables and equipment when usable equipment exists, stuck in a backroom or unnecessarily or redundantly plugged into a server.
3. Switch to Software-Defined Technologies
Software-defined data centers use virtualization to separate individual pieces of hardware — like a single server — into multiple virtual machines that can support different clients or projects.
For data center managers that already want to switch to SDN or virtualization, they may want to consider alternative network topologies — like leaf-and-spine architectures — that can work better with soft-driven data centers whose servers are running virtual machines and containerized software.
In data centers where the network has already been diagrammed, workers may have an easier time converting to an alternative topology, as all equipment is accounted for and it's possible to see how the existing topology can be rearranged at a glance.
Switching to a topology that better fits the needs of a data center can increase efficiency of resource use, which can lead to lower power costs and less physical footprint as all the resources offered by each piece of hardware are utilized.
Ensuring Data Center Efficiency
The increased demand for cloud-based computing resources had put new strain on data centers. Fortunately, there are practices that managers can implement to get the most out of their hardware, increasing efficiency without needing to buy new equipment.
Keeping stock of equipment, maintaining network diagrams and switching to software-defined technologies can all help managers improve the efficiency of their data centers, reducing energy costs and resource usage.
About Kayla Matthews
Kayla Matthews writes about data centers and big data for several industry publications, including The Data Center Journal, Data Center Frontier and insideBIGDATA. To read more posts from Kayla, you can follower her personal tech blog at ProductivityBytes.com.