Since the launch of ChatGPT in November last year,
the integration of business and artificial intelligence has been accelerating.
These workloads consume over five times more power than regular applications. Meanwhile, electricity rates have been raised five times from April to May last year, with further increases expected. As a result, the burden on companies for IT infrastructure operations is growing, driving greater interest in colocation services.
Performance Metrics Series to Enhance Your Data Center Evaluation Skills
Today, we will introduce N (Redundancy), a key metric for measuring data center availability.
Indicator "N" Representing Redundancy in Data Center
We will review part of the content from the "Comparing the Data Centers – Tier" segment.
What is data center redundancy?
Redundancy refers to the duplication of critical components in a system to ensure continuous operation, even if a failure occurs. In data centers, redundancy applies to power, cooling, fire suppression, and security systems, enabling uninterrupted IT infrastructure operation. Key components like power supply and cooling systems are duplicated to prevent downtime. Tier classification quantifies redundancy levels, helping businesses assess the resilience and fault tolerance of a data center. The ultimate goal of redundancy is to achieve zero downtime, ensuring business continuity and uninterrupted service.
Why Redundancy is Essential in Data Centers
The situations where redundancy is essential (where equipment operation must not be interrupted) are diverse. These include unpredictable events such as power outages, accidents, natural disasters, human errors during operations, equipment aging, and equipment failures. Additionally, there are planned situations that occur regularly, such as maintenance-related component replacements and test operations.
Backup to replace the interrupted equipment
N is a metric that indicates whether a backup is configured to replace interrupted equipment.
N represents the method and capacity of the backup.
There are two data centers, both of which are equipped with backup generators to prepare for power outages. Data Center A has a backup capacity to last for 10 minutes, while Data Center B has a backup capacity to last for 30 minutes. In this case, can we say that both data centers have the same level of redundancy? While their availability levels may be the same, their operational efficiency differs, which could impact the cost of using the data center. How can we distinguish between them in this context?
[IMAGE : DAOU]
Defining N
N refers to "Need", which represents the total capacity of Critical Load 1 and 2 (IT load). By definition, N does not include any redundancy, making it susceptible to single points of failure. On the other hand, N+1 signifies that a backup configuration is in place. If it refers to the backup of components (equipment), it means that additional capacity is available. If it refers to the backup of paths (Path or Distribution), it means that an alternate path is configured to ensure operational continuity.
1 Load : Total Power Consumption Capacity of a Data Center
2 Critical load or IT load :The power consumption of IT infrastructure (servers, storage, network, and security) within a data center.
Understanding N Through a UPS Example
Let’s take a closer look at UPS (Uninterruptible Power Supply) and batteries, which are key components of a data center's electrical infrastructure. UPS stands for "Uninterruptible Power Supply", which provides continuous power in case of power outages.
N configuration provides only the minimum capacity to support the critical load. For example, if the critical load is 10MW, an N configuration would use a single 10MW UPS. If this UPS fails, the entire system loses power, leading to a total outage.
N+1 redundancy adds one extra component to the required N components. For instance, if N equals 4 UPS units, an N+1 system would have 5 UPS units. If one UPS fails, the additional unit takes over the load, ensuring continuous operation. This design provides a balance of cost, energy efficiency, and system reliability. For greater protection, some data centers adopt N+2 redundancy, which includes two extra components. While this offers higher resilience, it also increases costs and system complexity.
In summary, N+1 is a widely used approach as it offers cost-effective redundancy while ensuring continuous uptime even during equipment failure.
How to choose the right redundancy configuration
As business and IT integration accelerates, data center redundancy becomes vital for business continuity. Choosing the right redundancy configuration depends on your IT environment, business goals, and budget. It's important to ensure the data center has the proper configuration, capacity, and operational capabilities for its electrical and mechanical systems. For the best approach, consult with an account representative, sales engineer, or solution architect. Our experts are ready to provide tailored guidance for your specific needs.
Content is protected by copyright law and is owned by Daou Technology Inc.
It is prohibited to modify or commercially use this content without prior consent.
Featured images via gettyimages.
References
1. .
2. .