Skip to content Skip to footer

How demand for data is shaping the evolution of data centre design

Share on facebook
Share on twitter
Share on linkedin

Brian Johnson, ABB’s global data centre lead, explores how unprecedented data demand is shaping the evolution of data centre design.

There is no doubt that the data centre sector is witnessing the highest level of demand in its short history. Driven mainly by our increasing thirst for content delivery networks for entertainment, education, home working and online commerce, data centre usage has increased 47% during recent months.

Demand is so high that the amount of data we now generate is growing 40 times as fast as the world’s population and we expect IoT devices to create 90 zettabytes (ZB) of data by 2025.

With the data centre construction market predicted to grow by $31.7 billion during 2020-2024, we will see increasing demand for more agile operations tailored to specific needs and the evolution of system design.

Modern data centres are large purpose-built, mission critical pieces of infrastructure, with much more in common with industrial facilities than with commercial ones. As such, it seems natural that industrial-grade automation systems should be used to monitor and manage data centre infrastructure and design, particularly as the impact of 5G and IoT starts to be felt.

Over recent years, the data centre industry has evolved from an IT support system to on-demand scalable services that deliver increasing levels of security and sustainability.

Once dominated by purpose-built on-premise enterprise data centres, we have seen a marked shift to off-premise colocation and cloud data centres, largely driven by the financial benefits of renting versus owning and higher levels of reliability. With a global spend of $38 billion on colocation services predicted by 2023 and multi-tenant data centres experiencing 5% growth, this is a trend that is set to continue.

These increasingly popular services for enterprise clients allow companies to easily rent colocation data centre space from third parties or utilise cloud data centres, eliminating the need for building, cooling and security infrastructure. They also reduce the need to manage IT components, such as servers, data storage and firewalls.

Coinciding with technological expansion, colocation providers are increasingly applying sound business practice to the design and construction of data centres. So much so, that data centres are now seen as critical profit centres, with strategies being aligned with business realities and capital and operational costs highly scrutinised.

The task of creating flexible fortress-like data centres that safely and securely store and manage business critical data and applications under every conceivable situation, while simultaneously accommodating both short-term and long-term growth, is daunting.

Flexible design

In response, the industry has had to evolve new concepts for design, construction and operation. Pressure for high levels of reliability and maintenance are central to a holistic approach, satisfying industry requirements for redundancy and fault tolerances.

To make sure that equipment replacement or removal does not impact critical load, data centres must be designed in a flexible and scalable way. This has led to broader standardisation of design to improve operational reliability.

Given current demand levels, speed to market will heavily influence how designs evolve. Fast track project execution, short delivery cycles and cost-effective utilisation of assets are key to the success for data centre projects, as they have a direct impact on capital costs.

In response to increasing demand for speed, we have seen a growth in pre-engineered solutions. These packaged solutions ensure safe, secure and continuous operation in a rapidly evolving data centre landscape.

With limited site-work and smooth start up processes, they offer flexibility, scalability and cost-effectiveness to ensure a short and effective startup, along with operative reliability and maintainability.

Conventional electrical topologies can be implemented in several different configurations, depending on project requirements and site conditions.

Topology for safe, smart and sustainable operations

Although most data centre electrification systems are unique, there are three main underlying topologies; system plus system, shared redundant and block redundant, all of which have different advantages.

System plus system topology

This uses two totally independent systems to feed the critical load and is the basis of design for more conventional data centres, such as enterprise and colocation companies. While they have a strong reliability record, these systems can be prohibitively high to operate with a maximum asset utilisation of 50%.

Shared redundant topology

This is typically used by colocation, hyperscale and cloud data centres to allow customers to improve utilisation by up to 66% and achieve lowest possible costs and peak power efficiency.

Block redundant topology

For data centres with single corded loads, block redundant topology, which uses a Static Transfer Switch (STS) to transfer critical load from the primary system to the reserve or catcher system, can achieve an asset utilisation of 80%.

This transformation from system plus system to more ‘modern’ block redundant topology is not only driven by the increasing scale of data centres, but also by improvements in networking.

Networks have shifted from previous monolithic on-premises enterprise to cloud-based Enterprise Resource Planning (ERP), which deliver faster deployments and immediate enterprise-wide availability, with the most up-to-date systems offering IoT and machine learning capabilities.

Evolution is key

Standard solution architectures, the pre-engineered integration of products for all electrical distribution needs from the utility service to the rack, have evolved as the answer to the data centre industry’s fast changing needs. They tackle increasing demand for flexibility, scalability and cost-effectiveness to ensure quick start-up, reliability and maintainability.

To keep pace with unprecedented levels of demand from our data hungry world, the way we design our data centres will have to continue to evolve. It is about protecting our future and is not an option, it is a given.

Show CommentsClose Comments

Leave a comment