This audio was created using Microsoft Azure Speech Services
There are trends and technologies making it more realistic for companies to forego building their own centralized data centers, and instead use the spare space, power, and cooling capacity of their distributed office buildings to house micro data centers to support their business. Using this approach of deploying single-rack data centers can save up to 42% capex, compared to building a new centralized data center. Let me explain.
The Main Drivers for Micro Data Centers
There are four main drivers for the adoption of micro data centers:
- Scalability
- Speed of deployment
- Reliability
- Outsourcing to the cloud and colocation
1. Scalability is a strategy to conserve capital by paying only what you need, when you need it. Think back to the lessons the data center industry learned during the “dot com” bubble. Micro data centers and their IT kit can more easily be “stepped and repeated” to accommodate growth of the business.
2. Speed of deployment is basically how fast you can procure and install a data center. The larger the data center, the more steps, people, and complexity are involved. The smaller the data center, the more simple it becomes to order a single part number that comes with everything you need, including the IT gear.
3. Reliability generally increases as standardization increases. Rule of thumb – the larger a data center (e.g. footprint, kW capacity), the more customized it becomes.
4. Outsourcing to the cloud and colocation are leaving companies with less equipment to house in a centralized data center. At the same time, the network equipment at the edge like offices and retail locations require added power and cooling reliability because the operations may be entirely dependent on connectivity to the cloud.
The IT Enablers Making Micro Data Centers Possible
Enablers are things that make it possible or easier to accomplish something. In this regard, there are three main IT technologies that enable the use of micro data centers. Hyperconverged infrastructure (HCI), compaction, and virtualization. These three technologies allow IT admins to deploy compute, storage, and networking in a single 2-3U appliance. Furthermore, the HCI management software simplifies everyday tasks like patching, spinning up new machines, provisioning storage, etc. The software also allows these distributed nodes to operate as a logical single data center. This is what ultimately allows the leap from the concept of breaking up a centralized data center into many smaller micro data centers, and distributing them throughout the many buildings available in a company’s real estate portfolio.
Probable Cost: Proving the Worth of Micro Data Centers with Capital Expense Analysis
Finally, cost. To quantify how much one could save by utilizing the sunk costs in existing buildings, my team developed a capital expense (capex) model that takes into account the physical infrastructure that exists in a building (like generator, core & shell, electrical gear, cooling system, fire suppression, etc). When all expenses are added up, you find that deploying micro data centers into existing buildings can save 42%. Barring outstanding expenses to consider like owning vs. leasing of the building, this capex savings is a strong variable to leverage when deciding on your IT architecture transformation.
To learn more about this research, watch our latest Meet the Expert video on the design, deployment, and management of micro data centers, or read White Paper 223: Cost Benefit Analysis of Edge Micro Data Center Deployments.
