DATA CENTER TIERS | Uptime – 1, 2, 3 & 4 Explained

This article provides a comprehensive understanding of the Data Center Tier Classification System, including the Uptime Institute and its role in classifying data centers, how important it is, explains who it is, and the services they provide, such as infrastructure, management and operation, strategic management, and sustainability.

🚀 Use the table of contents to move where needed, as this is a bit of a lengthy one.

---

Data center tiers, created by ‘the Uptime Institute’, are a standardized system of classification focusing upon ‘uptime and availability’ of critical IT infrastructure that ranges from Tier 1 [I] – Basic Capacity to Tier 4 [IV] – Fault Tolerant.

• Tier I – Basic Capacity
• Tier II – Redundant Site Infrastructure
• Tier III – Concurrently Maintainable
• Tier IV – Fault Tolerant Infrastructure

Referring to and using them for providing exact direction within the Owner’s Project Requirements [OPR] allows organizations to clearly define their overall expectations, design, construction, and maintenance of their data center facilities and projects, ensuring that they have the necessary infrastructure and resources to sustain operations for their use, clients or users inline with the expected downtime.

🟩 High-Level Tier Requirements Table

This table provides a high-level overview of the Tier requirements for each certification that are described in more detail below:

🟩 Who is the ‘Uptime Institute’?

Founded in the mid-1990s by Ken Brill and grown into a respected authority in the industry, the ‘Uptime Institute’ is a leading global provider of data center common standards and best practices [Data Center Tier 1-4 Classification System] to help designers and operators create and optimize their facilities, focusing on performance, efficiency, and availability.

Today, the Uptime Institute has a global presence with offices in North America, Europe, and Asia. The organization is headquartered in New York, USA, and is a subsidiary of The 451 Group, a technology industry analyst firm.

🟩 What Services do they Provide?

Uptime provides four main types of services for the Data Center Industry:

• Infrastructure Services,
• Management and Operation Services,
• Strategic Management Services,
• Sustainability Services.

🟧 Infrastructure Services

The ‘Uptime Institute’ provides comprehensive assessments to clients focusing on the physical infrastructure [mechanical, electrical, water, security, emergency power, etc] within their data center, certifying them to their ‘Tier’ system from the Design Stage to Operations.

The services include third-party Tier Certification for clients building new facilities from design to operations.

🟧 Management and Operation Services

By completing a comprehensive evaluation of a client’s current staffing, maintenance, operational practices, and management protocols; validating existing practices, and issuing recommendations for improvement, the ‘Uptime Institute’ can provide its stamp of approval based on its Management and Operations (M&O) Program to demonstrate a client’s operational excellence for their internal requirements or external stakeholders.

🟧 Strategic Management Services [SCIRA-FSI]

Financial sector organizations use the SCIRA-FSI assessment service to assess the operational resilience of their critical infrastructure.

This service helps organizations meet the growing regulatory requirements and manage the complexity of hybrid IT infrastructures, providing a standardized approach, based on input from over 20 premier financial institutions worldwide, to assess and mitigate outages.

By using this service, financial organizations will better understand their operational resilience and take steps to address any potential risks, which aids in the reduction of risk for downtime, increasing operational efficiency, and providing third-party validation of their operations to internal and external stakeholders.

🟧 Sustainability Services

With growing concern over the environmental sustainability of data center operations, regulators, customers, partners, and investors are looking for ways to show that the digital infrastructure they rely on is being designed and operated sustainably.

The Sustainability Service has been created to allow support and guidance to be provided by ‘Uptime Institute’ that organizations may require to become more environmentally aware and reduce the impact of their IT infrastructure and operations on the environment, reducing greenhouse gases, power, water usage and recycling of equipment.

🟩 Who uses ‘Uptime’ and the Tier System?

The tier system can be used by multiple types of people and organizations:

• Owners of Data Centers looking to build new facilities,
• Owners of Data Centers looking at expanding an existing facility,
• Data Center Operators looking for an evaluation of an existing facility,
• Data Center Customers, such as companies looking at utilizing a third-party data center, need to understand its design/resiliency/uptime.

🔗 To read more on obtaining Uptime Tier Ratings, see our article: UPTIME TIER RATINGS | The Simple Steps to Obtain them

🟩 What is meant by ‘Uptime’?

‘Uptime’ basically means ‘Availability’, and is the most critical aspect of data center operations, as it refers to the facility’s reliability and ability in maintaining the quantity of time a data center is available and online, being able to provide services to its customers without disruption [down time].

A high level of ‘Uptime’ or ‘Availability’ is essential to ensure that our mobile phones, computers, and connected devices are served via the cloud and networks with the internet, software programs, email, work files, financial transactions, e-commerce, location, texts, and calls etc.

🟩 What is the Data Center Tier Classification System?

The Data Center Tier Classification System, usually split into two, design and operational, has been developed to provide a consistent and standardized method for evaluating and classifying data center facilities and their infrastructure performance and uptime/availability. Infrastructure means critical mechanical, electrical, water, emergency power, and security systems.

This article will concentrate on the ‘Design’ Element

🟧 Design/Topology

Under the design classification, a facility’s potential infrastructure performance or uptime is evaluated based upon four levels of performance, referred to as Tier 1 – Basic Capacity | Tier 2 – Redundant Capacity | Tier 3 – Concurrently Maintainable | and Tier 4 – Fault Tolerant, with each higher tier incorporating all of the features of the previous one.

🚀 In the Tier rating system for data centers, it’s important to understand that the main source of power is not the utility power, but rather the generator system. The utility power is considered as an economical backup option, and any disruptions from it are not considered a failure.

🟧 Equipment, Components and Systems Relates to

The are many types of equipment and components that the Tiers relate to, below we list some of the most common ones:

◻️ Mechanical Equipment

◻️ BMS/Control Equipment

◻️ Electrical Equipment

◻️ Plumbing and Drainage Equipment

◻️ IT Equipment and Services

◻️ Security Equipment and Systems

Below we will talk about 'redundancy' - we have written an article covering tis in more depth: Data Center Redundancy | N, N+1, N+2, 2N & 2N+1 Explained 

🟩 Tier 1 – Basic Capacity

Tier 1 is the lowest tier rating, providing the most basic capacity level with no redundancy provided by the designed and installed cooling or electrical equipment.

Below are some notes regarding a Tier 1 data center:

• The most basic capacity,
• No redundancy is built into the system or its components,
• Expectation of 99.671% uptime for the year [possible offline for up to 29 hours],
• It would suit a small business owner who does not want to rely on the general building infrastructure for their IT systems, providing a limited amount of resilience.

🟧 General Requirements

Although it is the lowest in tier ratings there are still some expectations that need to be met, in order to be awarded the certificate:

• Sufficient capacity – there is sufficient capacity within the system to meet the system’s requirements.
• An uninterruptible power supply (UPS) for power sags – outages, and spikes, maintaining the overall service of the data center and servers.
• A dedicated area for IT systems – so no installing the servers in the corner of an office,
• Dedicated cooling equipment – that has the ability to run outside of office hours, so is not affected by time clocks and office user interference,
• Make-up Water – where evaporative cooling is being used, the makeup water should have an onsite storage facility of 12 hours for backup.
• An engine generator – for power outages, with a minimum of 12 hours fuel backup. Uptime also notes that a ‘fuel cell’ can be used.

🚀 The Engine Generator at [N] demand for this Tier CAN restrict the consecutive hours of run time based on the manufacturer’s requirements.

🟧 Tier 1 Data Center Graphic

The below shows a typical graphic of a Tier 1 Power and Cooling Infrastructure, these can differ from project to project dependent on the design and installation, so only provided as an example.

🟧 Benefits of Designing and Installing a Tier 1 System

The Tier 1 data center has some great benefits for a client, when compared to a facility that has not been designed inline with a standard and using general office infrastructure.

Below is a list of eight that we can think of:

🟧 Observations of Designing and Installing a Tier 1 System

If comparing against having the data center infrastructure running from a local office type power and cooling system, there are not that many observations, as the upgrade is positive.

The only items that we would see as negatives are:

🟩 Tier 2 – Redundant Site Infrastructure Capacity

Tier 2 data centers enhance the Tier 1 design by providing additional redundant critical power and cooling capacity/components to the system which will allow the facility to have a higher margin of safety and ‘uptime/availability’.

Below are some notes regarding a Tier 2 data center:

• Enhanced capacity over the Tier 1 design,
• Redundancy is built into the critical power and cooling components but not the distribution paths,
• Expectation of 99.741% uptime for the year [possible offline for up to 22 hours],
• It would suit small to medium business owners looking for a bit more protection from disruption of their IT system, especially when maintaining the critical equipment.

🟧 General Requirements

A Tier 2 Classified Data Center should include the following within its design, installation and maintenance, as a minimum:

• Components can be removed from service – When required via planned maintenance or operational issues, components can be removed from its service without impact on the critical environment it serves. Achieved by introducing redundancy into parts of the design.
• Single non-redundant distribution path – removing a distribution path from service for maintenance or operationally will generally require shutting down the critical environment it serves.
• Uninterruptible power supply (UPS) – power sags, outages, and spikes, maintaining the overall service of the data center and servers and allowing some maintenance with no impact to the data center operations,
• Redundant UPS capacity [N+1] – allowing for some operational failures and maintenance with no impact to the data center operations,
• A dedicated area for IT systems – so no installing the servers in the corner of an office,
• Dedicated cooling equipment – that has the ability to run outside of office hours, so is not affected by time clocks and office user interference,
• Redundant chiller capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant heat rejection capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant pump capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant cooling unit capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant chiller control capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Make-up Water – where evaporative cooling is being used, the makeup water should have an onsite storage facility of 12 hours for backup.
• An engine generator – for power outages, with a minimum of 12 hours of fuel backup during regular operation. Uptime also notes that a ‘fuel cells’ can be used.
• Redundant engine generator capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant generator fuel system capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,

🚀 The Engine Generator at [N] demand for this Tier CAN restrict the consecutive hours of run time based on the manufacturers requirements.

🟧 Tier 2 Data Center Graphic

The below shows a typical graphic of a Tier 2 Power and Cooling Infrastructure, these can differ from project to project dependent on the design and installation, so only provided as an example.

🟧 Benefits of Designing and Installing a Tier 2 System

Combined with the Tier 1 infrastructure, a Tier 2 data center will provide some additional benefits:

🟧 Observations of Designing and Installing a Tier 2 System

If a review is being conducted reviewing a Tier 1 design Vs a Tier 2 there are a few observations that would need to be considered and planned for:

🟩 Tier 3 – Concurrently Maintainable

Tier 3 data centers enhance the Tier 2 design by providing additional redundant critical power, cooling capacity/components, and multiple independent distribution paths [backbones] able to serve the critical spaces when one is unavailable. This allows each item that serves the critical spaces to be shut down for maintenance or removal from service, with no impact on the ongoing operations of the facility.

Below are some notes regarding a Tier 3 data center:

• Enhanced capacity over the Tier 2 design,
• Redundancy is built into all areas, the critical power/cooling components and distribution paths,
• The expectation of 99.982% uptime for the year [possible offline for up to 1.6 hours],
• It would suit medium to large business owners looking for protection from disruption of their IT system, especially when maintaining/switching distribution paths.

🚀 An ELECTRICAL POWER Backbone/Distribution Path is considered as the path from the on-site power generation system [generator/fuel cell], to the input of the critical UPS equipment serving the IT equipment or mechanical equipment.

🚀 A MECHANICAL Backbone/Distribution Path is considered as the pipework and infrastructure for transporting the chilled water from the critical areas via CRAH/CRAC or Fan Walls etc to the atmosphere via Pumps, Air Cooled or Water Cooled Chillers, and Cooling Towers [Chilled Water System and Condenser Water System].

🟧 General Requirements

A Tier 3 Classified Data Center should include the following within its design, installation, and maintenance, as a minimum:

• Components can be removed from service – Every component that provides the capacity to the IT critical areas can be removed from service with zero impact on the environment it serves,
• Concurrently maintainable – Every component that provides the capacity to the IT critical areas can be removed from service with zero impact on the environment it serves,
• Sufficient capacity – sufficient capacity is available for critical environments when any component is removed from service,
• Redundant distribution paths – removing a distribution path from service for maintenance or operationally will not require shutting down the critical environment it serves; this includes valving arrangements,
• Critical IT equipment dual powered – IT equipment is to be dual powered to ensure zero interruption during power failures. Can use transfer switches [STS/ATS],
• Uninterruptible power supply (UPS) – power sags, outages, and spikes, maintaining the overall service of the data center and servers and allowing some maintenance with no impact to the data center operations,
• Redundant UPS capacity [N+1] – allowing for some operational failures and maintenance with no impact to the data center operations,
• A dedicated area for IT systems – so no installing the servers in the corner of an office,
• Dedicated cooling equipment – that has the ability to run outside of office hours, so is not affected by time clocks and office user interference,
• Redundant chiller capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant heat rejection capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant pump capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant cooling unit capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant chiller control capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Make-up water – where evaporative cooling is being used, the makeup water should have an onsite storage facility of 12 hours for backup,
• Make-up water – systems to be concurrently maintainable,
• An engine generator – for power outages, with a minimum of 12 hours of fuel backup during regular operation. Uptime also notes that a ‘fuel cells’ can be used,
• Engine generator – to be rated for continuous usage,
• Redundant engine generator capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Engine generator concurrently maintainable – meets the requirements throughout and proven via testing,
• Redundant generator fuel system capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,

🚀 The Engine Generator at [N] demand for this Tier is NOT ALLOWED to have restricted consecutive run hours and restrict the consecutive hours of run time based on the manufacturer’s requirements.

🟧 Tier 3 Data Center Graphic

The below shows a typical graphic of a Tier 3 Power and Cooling Infrastructure, these can differ from project to project dependent on the design and installation, so only provided as an example.

🟧 Benefits of Designing and Installing a Tier 3 System

Combined with the Tier 1 and 2 infrastructure, a Tier 3 data center will provide some additional benefits, similar to those shown for the Tier 2:

🟧 Observations of Designing and Installing a Tier 3 System

If a review is being conducted between a Tier 2 design Vs a Tier 3, there are a few observations that would need to be considered and planned for:

🟩 Tier 4 – Fault Tolerant Infrastructure

Designed for the most complex and stringent protection of the IT critical infrastructure, providing the highest level of reliability/uptime. A Tier 4 data expands the Tier 3 design to encompass fault tolerance.

Incorporating ‘Fault Tolerance’ will allow the system to provide redundancy to every system, distribution path/backbone or component serving the IT spaces, meaning that if a single component or complete path fails or is taken offline for maintenance, there is no impact to the operations.

Below are some notes regarding a Tier 4 data center:

• Enhanced capacity over the Tier 3 design,
• Fault tolerance and redundancy is built into all areas, the critical power/cooling components and distribution paths,
• Expected 99.995% uptime for the year [possible offline for up to 26.3 minutes],
• It would suit a large complex business that is looking for the most stringent protection from disruption of their IT system.

🚀 An ELECTRICAL POWER Backbone/Distribution Path is considered as the path from the on-site power generation system [generator/fuel cell], to the input of the critical UPS equipment serving the IT equipment or mechanical equipment.

🚀 A MECHANICAL Backbone/Distribution Path is considered as the pipework and infrastructure for transporting the chilled water from the critical areas via CRAH/CRAC or Fan Walls etc to the atmosphere via Pumps, Air Cooled or Water Cooled Chillers, and Cooling Towers [Chilled Water System and Condenser Water System].

🟧 General Requirements

A Tier 4 Classified Data Center should include the following within its design, installation, and maintenance, as a minimum:

• Any Fault – any potential fault must be detected, isolated, and contained while maintaining ‘N’ capacity to the critical loads,
• Single failures – No single failure of a component, distribution path or system will impact the operation of the facility. This includes valving arrangements etc.
• Infrastructure ‘controls’ operation – upon failure, will provide an autonomous reaction to any failure, with no impact on the critical systems,
• Compartmentalized – Complementary systems and distribution paths should be physically isolated from one another. For example, this means separate chilled water systems serving dual coil CRAH’s.
• Components can be removed from service – Every component that provides the capacity to the IT critical areas can be removed from service with zero impact on the environment it serves,
• Concurrently maintainable – Every component that provides the capacity to the IT critical areas can be removed from service with zero impact on the environment it serves,
• Sufficient capacity – sufficient capacity is available for critical environments when any component is removed from service,
• Redundant distribution paths – removing a distribution path from service for maintenance or operationally will not require shutting down the critical environment it serves,
• Critical IT equipment dual powered – IT equipment is to be dual powered to ensure zero interruption during power failures. Can use transfer switches [STS/ATS],
• Uninterruptible power supply (UPS) – power sags, outages, and spikes, maintaining the overall service of the data center and servers and allowing some maintenance with no impact to the data center operations,
• Redundant UPS capacity [N+1] – allowing for some operational failures and maintenance with no impact to the data center operations,
• A dedicated area for IT systems – so no installing the servers in the corner of an office,
• Dedicated cooling equipment – that has the ability to run outside of office hours, so is not affected by time clocks and office user interference,
• Redundant chiller capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant heat rejection capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant pump capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant cooling unit capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Redundant chiller control capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Make-up water – where evaporative cooling is being used, the makeup water should have an onsite storage facility of 12 hours for backup,
• Make-up water – systems to be concurrently maintainable,
• An engine generator – for power outages, with a minimum of 12 hours of fuel backup during regular operation. Uptime also notes that a ‘fuel cells’ can be used,
• Engine generator – to be rated for continuous usage,
• Redundant engine generator capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,
• Engine generator concurrently maintainable – meets the requirements throughout and proven via testing,
• Redundant generator fuel system capacity [N+1] – allowing for some operational failures and maintenance with no impact on the data center operations,

🚀 The Engine Generator at [N] demand for this Tier is NOT ALLOWED to have restricted consecutive run hours and restrict the consecutive hours of run time based on the manufacturer’s requirements.

🟧 Tier 4 Data Center Graphic

🟨 Power System

The below shows a typical graphic of a Tier 4 Power Infrastructure, these can differ from project to project dependent on the design and installation, so only provided as an example.

🟨 Cooling System

The below shows a typical graphic of a Tier 4 Cooling Infrastructure, these can differ from project to project dependent on the design and installation, so only provided as an example.

🟧 Benefits of Designing and Installing a Tier 4 System

Combined with the Tier 1, 2 and 3 infrastructure, a Tier 4 data center will provide some benefits similar to those shown for the Tier 3:

🟧 Observations of Designing and Installing a Tier 4 System

If a review is being conducted between a Tier 3 design Vs a Tier 4, there are a few observations that would need to be considered and planned for, similar:

---

⬜ Definition of Terms

Below is a high-level list of terms that are included within this article for reference in case you are unsure what they mean:

Tier – A level of performance that a data center is categorized into based on its infrastructure’s design and operational characteristics.

Uptime – The availability of a data center’s systems and the time it is operational and available for use, also known as ‘availability’.

Data Center – A facility that houses critical IT systems and equipment, including servers, storage, and networking equipment.

Basic Capacity – Refers to the lowest level of performance in the tier classification system, including the minimum design and operational requirements for power and cooling systems.

Redundant Capacity – Refers to the duplication of critical components within a data center infrastructure to ensure continued operation in the event of a failure or outage of components.

Concurrently Maintainable – The ability to maintain and upgrade system components and equipment without impacting the overall operation of the data center.

Fault Tolerant – The ability of a system to continue functioning, even if one or more components fail.

Infrastructure – The physical components of a data center, including power, cooling, servers, storage, and networking equipment.

Operations – The day-to-day management and maintenance of a data center, including monitoring, maintenance, and upgrades to the infrastructure.

---

⬛ Related Articles

⬛ Latest Articles

---

⬛ Author

If you would like to know more about the author, here is my Personal Linkedin Account