Data center costs are a major line item in every company's IT budget. But where do these investments really go, and how can business leaders optimize their infrastructure spending? 

To help ensure that your business is getting a good ROI rather than squandering resources, you need to first have a full understanding of data center costs, types and architectures. 

What are data centers?

Data centers house business-critical backend IT infrastructure and data storage systems.

A typical data center architecture has three main elements:

  • Computing hardware systems, such as mainframes, servers, hard drives and databases
  • Supporting data center equipment, such as switches, power supplies, cooling systems, etc.
  • Logical network layer, which establishes connectivity between the data center equipment and other business systems

Combined, this type of technical setup acts as a central repository of corporate data and allows it to be further processed and distributed across the organization. 

Key considerations for a data center 

  • Uninterrupted power supply and backup energy sources
  • Extra servers/sites available for redundancy and backups
  • High-speed network connectivity and sufficient bandwidth
  • Located in a facility with temperature and humidity controls
  • Adherence to industry-recognized standards such as ANSI/TIA-942
  • Proper physical and network security assets
  • On-site or on-call engineers, available for maintenance
  • True total cost of ownership (TCO) for data centers, including operating and maintenance expenses

Types of data centers

In 2021, businesses are projected to spend more than $200 billion on global data center infrastructure, a jump of 6 percent from last year. Much of the spending will be directed at new hardware purchases — servers and supporting hardware solutions — as well as legacy assets modernization.

Cloud infrastructure spending is accelerating as well, with more and more leaders completing cloud migration projects and settling into the new routines of running workloads in a multicloud environment. According to Synergy Research Group, enterprise spending on cloud infrastructure increased by 35 percent last year, whereas spending on on-premises data centers declined by 6 percent.

Without doubt, many market leaders today have a growing infrastructure footprint, spanning over several on-premises interconnected data centershybrid cloud setups and virtualized data center facilities located in different geographic locations.

Each option has its merit and competitive strengths compared to others.

On-premises data centers

An on-premises data center is a dedicated room, featuring server racks and supporting equipment, located on the same grounds where most business users operate.

Today, on-premises remains a data center configuration that most businesses (whether they use the cloud or not) rely on at least to some extent. Many organizations choose to keep core systems and data on-site for security and compliance reasons, as well as to minimize latency for bulk data transfers. 

Pros of on-premises data centers:

  • Offer single tenancy to meet regulatory requirements
  • Purpose-built, customizable systems that are optimized for specific workloads
  • Full control and complete data visibility

Limitations of on-premises data centers:

  • Aging infrastructure and higher overall CAPEX spending
  • In-house talent required for support and maintenance
  • Long procuring cycle for obtaining extra capacities

Colocation data centers  

The colocation model enables companies to rent a portion of a data center facility from another vendor for a monthly or annual fee. The vendor provides a suitable physical space for placing servers and implements the necessary security, environmental and connectivity controls. As a customer, you are mostly responsible for bringing, managing and maintaining your own hardware (although some providers also offer hardware for lease). 

Pros of colocation data centers:

  • Reduced CAPEX costs
  • High degree of hardware customization
  • Can act as a backup site
  • High uptime availability, backed by SLAs
  • Full control over data center configurations and management

Limitations of colocation data centers:

  • Long-term colocation contracts may be less effective than on-premises
  • Harder accessibility for maintenance personnel
  • In some cases, not all providers will meet the security and compliance regulations of your industry

Virtualized data centers 

Data center virtualization is a middle-ground between complete cloud migration and on-premises storage. It's a common approach to creating private clouds and operating hybrid data centers — a setup where some assets are hosted locally, and others reside in on-site or managed private cloud(s) and public clouds.

By using virtualization products and software-defined solutions, you can create a highly available, scalable and secure data storage architecture hosted on-premises.

Pros of data center virtualization: 

  • Data center consolidation leading to hardware savings
  • On-demand access to extra computing or storage resources
  • Reduced energy consumption
  • Faster provisioning of extra capacities
  • Improved data mobility
  • Can support disaster recovery (DR) plans

Limitations of data center virtualization: 

  • Extra implementation and software licensing costs
  • Present a single point of failure
  • Requires careful planning and resource usage orchestration

Cloud data centers 

Cloud infrastructure and Infrastructure as a Service (IaaS) cloud services usage have swelled over the past year. As Gartner predicted in 2019, many organizations opt to fully transfer their data centers to the cloud or opt for colocation. "By 2025, 80 percent of enterprises will shut down their traditional data centers," stated Gartner. "In fact, 10 percent of organizations already have as of 2019."

Cloud services providers provide access to their physical infrastructure, packed with best-in-class Cisco or Intel hardware and hardened for utmost security, 99.99 percent availability and unbeatable fault-tolerance. As a cloud data center user, you commission remote, on-demand access to computing and data storage resources. 

Cloud data centers offer a more attractive pay-per-use cost structure, as well as just-in-time two-way scalability — the two main attractors for growing enterprises. 

Pros of cloud data centers:

  • Switch from CapEx to OpEx costs
  • On-premises resources can be retired or re-assigned as needed
  • Access to virtually unlimited capacities
  • Rapid up-scaling and down-scaling
  • Higher global availability due to access to global data centers
  • Freeing up IT staff who were responsible for data center maintenance

Limitation of cloud data centers: 

  • Knowledge required for right-sizing cloud infrastructure and optimizing spending
  • Data center security is a shared job — you need to ensure proper safeguards
  • Low to no possibility of hardware customization

If you opt for a public cloud, server access is shared with other tenants. Managed private clouds, on the other hand, allow customers to rent private infrastructure and prevent any co-mingling. 

Conclusion

When considering which data center type is best for your business, it's important to consider the challenges and opportunities that come with each option. WWT can help your company optimize your data center architecture through best practice recommendations and use case discussions. 

Need help with data center and IT infrastructure design?
Request a facilities infrastructure briefing