Data centre redundancy ensures uninterrupted operation by duplicating key components such as power supplies, servers and cooling systems. Configurations such as N, N+1, 2N or even 3N2 offer different levels of redundancy and an optimised level of security and stability. This redundancy means that data centres can operate with complete peace of mind, knowing that their infrastructure is protected against potential failures.
Maintaining uptime through redundancy
In the field of electrical infrastructure, data centre redundancy has always been the most effective method of increasing power availability and, consequently, service availability. According to reliability theories and experience, adding a redundant component makes the system more reliable.
The idea is simple: in a redundant system, if one component fails, the other continues to keep the system running smoothly.
Data centre redundancy involves duplicating critical components to prevent service interruptions. This is done is several ways:
- Hardware redundancy: duplication of servers, hard disks and other hardware.
- Power path redundancy: multiple electrical circuits to provide a continuous power supply.
- Network redundancy: multiple network links.
These strategies ensure high service availability.
The Uptime Institute classifies data centres into four levels (Tier I to IV), each offering an increasing degree of redundancy and reliability. Let’s look at the different levels of redundancy
Levels of redundancy: Tier 1, Tier 2, Tier 3 and Tier 4
Data centre redundancy levels vary depending on their Tier classification.
Tier 1 is the most basic level, with a single power supply and cooling system. There is no redundancy, which entails approximately 28 to 29 hours of downtime per year.
Tier 2 is partial redundancy, which includes redundant components such as emergency generators and emergency cooling equipment. Availability is improved.
Tier 3 offers complete redundancy. Each critical component, whether it be the power supply or cooling, has N+1 redundancy. This means that an additional component is available for each essential component. This configuration provides an uninterrupted power supply, with an availability rate of 99.982%, or around 1.6 hours of downtime per year.
Tier 4 data centres offer 99.995% availability, or around 26 minutes of downtime per year. This redundancy ensures exceptional fault tolerance. Each critical component is fully redundant with a 2N+1 configuration.
Redundant power supplies in data centres
The components of a good power supply
To maintain a reliable power supply in a data centre, several components are essential.
Generators are critical. They take over in the event of a power outage. These generators are often powered by diesel engines and need to be regularly tested to ensure that they work in an emergency.
Uninterruptible power supplies (UPS), also known as inverters, play a crucial role in guaranteeing back-up power during power outages.
Static transfer systems (STS) manage the transfer from one power source to another without interruption. They switch instantly to a backup source in the event of failure of the main source.
Power distribution units (PDUs), which distribute electricity to the various data centre equipment.
UPS: A key element of power redundancy
UPS (Uninterruptible Power Supplies) systems step in as soon as a power cut occurs, ensuring uninterrupted operation of critical equipment.
N+1 and N+X configurations are commonly used to improve redundancy. In an N+1 configuration, an additional UPS is added for each group of UPSs, while N+X allows several redundant UPSs to be added.
UPSs generally operate in double conversion mode, transforming alternating current into direct current and vice versa, thus stabilising the voltage supplied to servers to protect loads.
Delphys XL is a high-performance UPS solution specially designed to secure the most critical applications. It offers:
- exceptional intrinsic protection
- a unique brick concept that eliminates any single point of failure
- a solution suited to all data centre architectures, with each power brick operating independently, ensuring distributed control.
This UPS features an innovative operating mode: the Smart Conversion mode.
Smart Conversion mode is based on an advanced algorithm
that constantly monitors network quality and selects the optimum operating mode between Double Conversion (VFI) and Line Interactive in real time.
In the event of a disturbance on the grid, the UPS instantly switches to double conversion mode with a transfer time of 0ms in compliance with the Class 1 requirements of standard IEC 62040-3..
This mode reduces losses by a factor of 5 and saves 350 MWh of energy per year with no risk to power continuity.
2N, 3N2 and Catcher redundancy: What is it?
2N: Definition and benefits
2N redundancy means that a data centre has twice the necessary quantity of each critical component. This configuration ensures that no single point of failure can disrupt overall operation.
There are many advantages to this architecture. Firstly, it offers exceptional reliability. Even in the event of a component failure, the system continues to operate without interruption.
However, to deliver on its promises, this electrical design needs all the electrical equipment (generators, inverters, UPS, switches, etc.) to be redundant, which means investing in twice as much equipment.
3N2: Definition and benefits
Distributed architectures, such as ‘4N3’ or ‘3N2’, aim to optimise power redundancy by sharing it between different systems. In this configuration, out of a total of four systems, only three are needed to power the load. This means that there is always a spare component for each pair of units in operation.
The benefits are clear: It optimises the implementation of UPSs and reduces investment. Unfortunately, at the cost of complexity. This architecture requires all the UPS to be installed beforehand, which imposes cabling constraints and limits its compatibility with the modularity requirements of data centres.
Catcher: Definition and benefits
The Catcher architecture effectively creates an N+1 or N+2 architecture within the UPS while maintaining fault tolerance and the possibility of simultaneous maintenance thanks to the use of static transfer systems (STS) which are placed between the UPS and the load. STS units are used in this configuration to:
- To transfer critical load from the main or active system to the Catcher,
- To isolate in the event of a short circuit.
Downstream of the STS units, the electrical distribution system can be designed in a similar way to a 2N architecture.
With this configuration, a UPS can operate at a load of 75% or more, while the Catcher remains unloaded under normal conditions.
Catcher architecture is currently used by large and medium-sized data centres, including cloud hosting and colocation facilities, as an alternative to traditional 2N architecture. This approach offers a similar level of availability while being more efficient and less costly in terms of capital.
The Catcher model stands out for its ability to optimise redundancy while limiting investment costs. Unlike the 2N and 3N2 configurations, the Catcher model uses a flexible approach, making it easier to adapt to the specific needs of data centres. This flexibility is particularly advantageous for expanding facilities.
There are many advantages to the Catcher model:
- Cost control: Fewer redundant components needed, which reduces initial costs.
- Better UPS sizing
- Simplified maintenance: Modules can be replaced individually without interrupting service.
Example: With a Catcher architecture, a 1MW room will require a 1MW UPS upstream and an STS of around 1600 amps. In the event of a UPS failure, this STS will transfer the load to a spare UPS or Catcher, which will also serve as redundant equipment for other rooms.
By adopting this model, businesses can guarantee high availability of their services while keeping costs under control.
Role of the static transfer switch system
Static transfer systems (STS) allow critical load to be transferred from a failed power source to an alternative source without interruption.
Unlike ATS, the STS uses semiconductors, such as thyristors, to switch between two power sources. This enables virtually instantaneous switching, in just a few milliseconds. This speed is essential for critical applications that do not tolerate even brief interruptions to the power supply. As a result, the STS is particularly well suited to sectors where continuity of power supply is paramount, such as banking, finance, healthcare or data centres.
STS can also be integrated directly into data centre racks. They offer a compact and efficient solution for power management. Companies can therefore guarantee the reliability of their infrastructure while optimising the available space.
“STS technology makes it possible to achieve high levels of power availability while keeping costs under control,”
Xavier Mercier – Marketing Director EMEA at Socomec
Focus on STATYS - Socomec’s static transfer system
In a context where power supply continuity is a key factor in remaining competitive, Socomec’s STATYS static transfer system is particularly relevant.
With more than 35 years of expertise and millions of hours of use, Socomec is constantly improving its products and services. The fourth generation of STATYS guarantees uninterrupted power supply availability for applications ranging from 32 to 1800 A.
This range is specially designed for environments where the network cannot tolerate any interruption.
- The STATYS static transfer switch guarantees maximum resilience for total power availability, meeting all integration requirements.
- Microcontroller redundancy, physically separated for increased security.
- SCR driver with independent, redundant power supplies.
- Redundant cooling with a fan failure monitoring system,
Over 8,000 units are currently in operation around the world.
Want to find out more? Don’t hesitate to contact our team by filling in this form.
