dcsimg

Data Center Tiers: Formulating a Strategy

SHARE
Share it on Twitter  
Share it on Facebook  
Share it on Linked in  
Email  

Data centers serve as the foundation for the modern enterprise. They power the IT systems that run complex businesses, including servers, backup systems, telecommunications equipment, and numerous other technology components. More than 8.5 million data centers exist worldwide, and the figure in the U.S. now tops 3 million.

Keeping systems up and running is paramount. However, not all IT infrastructure is created equal — and not all business needs are the same. That’s why data center tiers exist. The term refers to different types of systems, components and infrastructure that are arranged into groups, or tiers.

Each tier is designed to address specific IT and equipment requirements. Tier 1, for instance, involves basic infrastructure requirements, while Tier 4 is comprised of the most complex components. Let's look at each of the Tiers.

Data Center Tier Standards and Requirements

Understanding how data center tiers work is crucial to designing an effective IT strategy.  The tier classification system was created by the Uptime Institute in the mid 1990s. Since then, the framework has evolved from a shared industry terminology to a global standard that includes third-party validation of data center critical infrastructure.

The tiering system is progressive, meaning that each tier is dependent on the tier below it and incorporates the requirements in the lower tier.

Here is the role that each tier plays in the data center:

Tier I. Basic Capacity

A Tier I facility incorporates dedicated site infrastructure to support information technology beyond an office setting. It can be thought of as a tool for achieving a tactical level of operational sustainability. The Uptime Institute notes that an organization’s Tier I needs are primarily driven by time-to-market and first-cost issues rather than lifecycle costs. It typically includes:

  • Dedicated space for IT systems, such as servers and backup devices.
  • Uninterruptible power supply (UPS) devices to handle power fluctuations and failures.
  • Cooling equipment to keep systems running at optimal temperatures.
  • Engine/power generators to keep systems operating and online during extended power outages.

Tier II. Redundant Capacity Components

A Tier II data center framework encompasses redundant critical power and cooling components. These provide a margin of safety from power disruptions and other major events that can force system shutdowns or cause damage. Tier II is also tactical in nature. Most organizations that rely on Tier I and Tier II capabilities don’t require real-time capabilities, according to the Uptime Institute. Tier II systems and devices typically include:

  • Power and cooling equipment such as UPS modules, chillers or pumps.
  • Engine generators.

Tier III. Concurrently Maintainable Systems

Tier III facilities can operate without impact to IT systems during equipment upgrades, changes or maintenance. A data center that boasts Tier III capabilities avoids shutdowns by establishing redundant delivery paths for power and cooling. Uptime Institute notes that Tier III (and Tier IV) site infrastructure solutions have an effective life beyond the current IT requirement. What’s more, they are typically utilized by organizations that recognize a cost of a disruption — in terms of actual dollars as well as market share.

Tier IV. Fault Tolerance

The infrastructure in a Tier IV data center builds on the capabilities of a Level III facility by building fault tolerance into the topology of the site. Fault tolerance describes the ability to limit the effects of a disruption or interruption before they reach IT operations. It is particularly important for mission critical applications and systems. This tier represents the most strategic level of protection. It often involves multiple redundant systems.

Data Center Tiers at a Glance

Tier

Overview

Key Components

Uptime

Risk

Tier I

Tactical. Has a single path for power and cooling, few redundant and backup components.

UPS modules, cooling equipment, power generators.

99.671% (28.8 hours of downtime annually).

Medium to high.

Tier II

Tactical Has a single path for power and cooling with some redundant and backup components.

UPS modules, chillers, pumps, power generators.

99.741% (22 hours of downtime annually)

Medium.

Tier III

Strategic. Has multiple paths for power and cooling and systems so that upgrades, updates and maintenance can take place without taking the data center offline.

UPS modules, chillers, pumps, power generators (more extensive deployment).

99.982% (1.6 hours of downtime annually).

Medium to low.

Tier IV

Strategic. Completely fault tolerant with redundancy for every component.

UPS modules, chillers, pumps, power generators (most extensive deployment).

99.995% (26.3 minutes of downtime annually).

Low.

The Evolution of Data Centers

As data centers have become more complex and interconnected — and as high-performance computing (HPC) workloads and cloud computing have moved into the mainstream of the enterprise — tiering strategies have had to address new and sometimes different issues. Make no mistake, modern facilities don’t resemble the data centers of the 1980s, which often housed mainframe or midrange super computing systems.

In order to adopt the right data center tiers, it’s particularly important to examine how a service provider has built and maintained its data centers. There are several key factors, and, in some cases, these may intersect with industry groups and standards, such as those created by the Telecommunications Industry Association (TIA). Key factors may include:

  • Geographic location/risk and earthing standards
  • Commercial building standards.
  • Cabling standards.
  • Cooling systems.
  • Fiber-optic coding.
  • Operational practices.
  • Equipment standards, including maintenance and replacement of UPS units, generators and other devices.

Data Center Tiers and Cloud Computing

As organizations look to build out a more agile and flexible IT framework, colocation and various cloud computing facilities — including hybrid clouds — require close scrutiny and different considerations. Indeed, today's data centers are often one large private cloud.

  • These facilities must deliver the level of availability businesses require — typically across multiple locations. Given the growth of multicloud computing, these multiple locations often entail a multicloud strategy.
  • Because an organization doesn’t have control of systems in the data center, service level agreements (SLAs) are a key consideration.
  • The Uptime Institute has expanded its tier ratings to include operational sustainability. This addresses staff and operational components, including prioritized behaviors and risks. It’s an aspect that businesses should consider as they move to cloud platforms and possibly Infrastructure-as-a-Service (IaaS).

How to Formulate a Data Center Strategy

A starting point for navigating data center tiers is to understand your organization’s needs and how a tiering strategy fits in — particularly as the march to the cloud accelerates. As businesses turn to outside data center providers for services, and in some cases use cloud companies, the challenges of vetting vendors grow.

Consequently, many organizations are turning to outside consultants and organizations to assess data centers based on their adherence to specific tiers. However, the Uptime Institute is the only organization that will rate and certify a data center. This ensures that a facility was constructed as designed, that site functions and equipment work as billed, and that the data center can demonstrate performance standards that match its tier designation. However, it’s also important to note that some companies now boast standards that exceed Uptime Institute and TIA.

It’s also important to recognize that tiering isn’t the only consideration in establishing or choosing a data center provider. Configuration management, security, utilization levels, licensing requirements, disaster recovery and, possibly, rack capacity and density issues are also important considerations. For business looking to tie in data center tiering, it’s also essential to understand the specific services offered and what quality of service (QoS) levels are guaranteed.

Nevertheless, a focus on data center tiers can help an enterprise boost performance, trim costs and ensure that the business supports current and future technology and performance requirements. According to AFCOM’s 2018 State of the Data Center Industry study, the need for diligence is growing. Nearly 60 percent of companies currently own between two and nine data center facilities but the average number of data centers that organizations manage will grow to 10.2 by 2021. Moreover, many existing facilities will require renovation and changes.

To be sure, as organizations grow and evolve — and business-critical services become mandatory — executing on a data center tiering strategy is increasingly important. It’s at the center of successful IT and business.

Key Steps to Improving Your Tiers

It’s important to match your organization’s requirements with appropriate data center tiers. Here are some tips for determining what your enterprise requires and establishing appropriate data center tiering levels:

Step 1: Assess Your Enterprise and your requirements.

Understand the specific availability and performance requirements of various enterprise systems and applications. For example, a branch office at a financial services firm doesn’t require the same level of performance and protection as an e-commerce system or a securities trading system. It may be necessary to extend this assessment to supply partners and customers.

Step2: Identify Risk Levels.

What types of disruptions or interruptions are possible? How will an interruption or system failure impact the business? What will it cost in terms of money and reputation for a system or application to go down for a few minutes or a few hours? It’s critical to document the costs of a business interruption and compare it to the cost of a maintaining or purchasing a specific data center tier.

Step 3: Assess Internal Needs and/or Outside Providers.

Determine what changes you will need to make if you are operating your own data center. If you are looking to an external provider for compute, bandwidth or cloud resources, determine the vendor’s certifications, rankings, policies, procedures, protections and what SLA they offer (including compensation for failure to meet standards).

Step 4: Map Systems to Tiers.

With all the necessary information on hand it’s possible to determine what makes sense — and what maximizes dollars — while reducing risk. It’s vital to avoid ambiguous language and decision-making and ensure that all tiering decisions are based on concrete and defensible criteria. It’s also paramount to build in notification systems to know if and when a data center doesn’t perform to established tiering criteria.  

Data Centers Tier Strategy: The Final Frontier

The most important step in your data center tiering strategy? Understanding exactly how important it is to have a fully developed strategy.

A focus on data center tiers can help an enterprise boost performance, trim costs and ensure that the business supports current and future technology and performance requirements. According to AFCOM’s 2018 State of the Data Center Industry study, the need for diligence is growing.

Nearly 60 percent of companies currently own between two and nine data center facilities but the average number of data centers that organizations manage will grow to 10.2 by 2021. Moreover, many existing facilities will require renovation and changes.

To be sure, as organizations grow and evolve — and business-critical services become mandatory — executing on a data center tiering strategy is increasingly important. It’s at the center of successful IT and business.



NewsletterDATAMATION DAILY NEWSLETTER

SUBSCRIBE TO OUR IT MANAGEMENT NEWSLETTER