Although there is no guarantee that large data centers are more efficient than small ones, large facilities have...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
two key benefits: scale and tools.
Business managers constantly seek ways to improve efficiency and boost performance to maximize data center returns.
Some organizations gain efficiency and save money via IT automation and management technologies, but data center size determines the usefulness of tools like DCIM. There are several questions to ask about efficiency investments' implications.
Big enough for DCIM
Now that data centers are a critical business asset, building management and IT systems management teams must pay close attention to managing the facility with fast response times to problems and changes.
Data center infrastructure management (DCIM) software tools put facilities management -- such as temperature and humidity monitoring -- with server and storage monitoring and other IT tasks on a single pane of glass. Administrators can provision servers in the optimal load-balanced spot to ensure safe room temperature, guard against unauthorized data center access, monitor energy efficiency and so on.
There is no single point of inflection that defines when or where DCIM should be deployed.
One meaningful measure of DCIM is work-versus-reward analysis. DCIM adds physical sensors and a layer of software to the environment, which adds to the workload for IT staff. Each business that evaluates DCIM must determine if the benefits of superior capacity planning, faster response to problems, lower energy use and so on outweigh the additional work needed to deploy DCIM sensors and software. Your organization might not be large enough for the additional effort and expenditure to pay off in tangible savings, or you might lack the talent and time in-house to make use of the tool set.
Also consider future cloud or outsourcing initiatives. While DCIM is appropriate for large cloud or hosting providers, organizations that plan to move workloads to cloud or to outsourcing providers generally reduce their dependence on owned data centers. Look for a provider with a detailed DCIM portal.
Virtualization versus data center size constraints
Generally, virtualization lowers space, power and cooling requirements, delaying new data center builds.
Hosting multiple workloads on one server vastly improves hardware use and simplifies tasks like workload migration and data protection. Since virtualization allows one server to do the work of 10, 15 or more physical systems, the number of systems needed to operate an enterprise, even a small business, drops significantly.
If virtualization allows for an average of 6-to-1 consolidation, an enterprise with 480 physical servers and workloads would only need about 80 servers -- not counting a handful of servers for backup and other ancillary tasks. A data center can do the same work with far less space, requiring less power and cooling.
The abstraction layer of virtualization also allows organizations to evaluate outsourcing alternatives like third-party infrastructure as a service (IaaS) or cloud providers. Outsourcing workloads prohibits data center growth and can stretch refresh cycles.
Once workloads are consolidated onto virtual servers, the data center facility may be too big. Use containment or partition off areas to reduce the air volume that must be cooled. For long-term results, consider renovating the unused space, or opt for more efficient on-demand builds like containerized data centers.
A tale of two data centers
A large company with 1,000 physical servers and a sizable IT staff can deploy virtualization and consolidate servers. They justify management tools like DCIM to measure power usage effectiveness, tech refreshes to get servers with higher operating temperatures, and tools that migrate workloads and power down lighter-used servers automatically to optimize energy use.
The cost and IT talent needed to deploy all of those technologies may not be practical for small businesses. Consider a company with 10 servers in a data closet. They may virtualize and consolidate some of the 10 servers, but the cost -- and the time training a small IT staff -- to deploy automation and advanced management tools like DCIM far exceeds the energy savings.
The large data center and small company can both deploy a hypervisor like VMware vSphere, Microsoft Hyper-V or Citrix XenServer and perform a 10-to-1 consolidation project. The large company will take 900 servers offline, while the small company will take out only nine, so the savings are much more apparent for the large data center. The large company's savings can be rolled into management tools that automate important tasks and streamline workflows, but the savings for the smaller data center aren't enough to buy management tools -- or pay for IT personnel to operate the tools.
Ultimately, smaller data centers can realize measurable efficiency improvements from well-established technologies like virtualization, but typically lack the capital to pursue enterprise-class management tools or initiatives that extract maximum level of efficiency.