Data at the edge

Mark Turner at Pulsant highlights the ways that the edge data centre market is being impacted by sustainability concerns, AI adoption and cloud repatriation

Edge data centres have witnessed a boom of interest, currently experiencing growth of 92% year on year. Edge data centres are small data centres that transmit and collect data close to the end-user and their devices, offering low levels of latency which can make for a better user experience.

As businesses across sectors aim to optimise their processes to adapt greener outputs, implement AI into their systems and are seeking cloud repatriation. These three goals that many businesses share contribute to the interest in edge data centres.

Edge data centres and sustainability

Data centres by their nature, consume high amounts of energy. With the rise of AI this increases the energy demand further because in order for businesses to adopt AI, they have to be able to store and analyse data to be able to train machine learning (ML) and if they utilise large language models (LLMs), this increases the energy demand further.

Looking at GPT-3 alone, producing it requires 1,287 megawatts hours of electricity and results in 552 tons of CO2 emissions.  This is the equivalent to 123 gasoline-powered passenger vehicles.

Governments across the globe are increasing their concerns around data centre construction because of the significant energy levels they produce which conflicts with national climate objectives and electricity grids. In Ireland, the national electricity demand is forecasted to reach 32% by 2026. This encourages data centre operators to look at more sustainable alternatives, that can help them reduce their carbon footprint, while they continue to address the market demands.

One of the ways in which edge data centres can become more sustainable is by implementing heat reuse strategies. Since they operate locally to buildings and public infrastructure, these could be heated by the heat generated from the data centre’s IT equipment. 

Today, edge data centre operators are employing some of the following strategies to try and reduce their PUE (power usage effectiveness) and associated carbon footprint.

Utilising energy efficient technologies. There have been investments made by edge data centre operators into heavily energy-efficient technologies and practices.  The technologies being invested in, are the likes of energy efficient hardware and advanced cooling systems that help reduce energy waste. Implementing AI technologies also presents an opportunity to optimise energy usage within data centres by dynamically adjusting cooling systems, server settings and scheduling workloads to take advantage of renewable energy sources.

Investing into renewable energy solutions. Many data centres operators are switching to renewable energy sources such as solar, wind and hydroelectric energy to power their facilities. Not only does this allow them to reduce their carbon emissions but it also allows them to obtain long-term stability in energy costs as renewable energy becomes more cost competitive.

Developing greener building standards. The design and construction of data centres are also being influenced by sustainable considerations as companies across different sectors are doing their bit to reduce their carbon footprint. Green building standards such as Leadership in Energy and Environmental Design (LEED) are being applied to data centre construction projects, which focuses on factors such as energy efficiency, water conservation and promotes the use of environmentally friendly materials.

Pledging to become carbon neutral. Data centre providers are committing to achieve carbon neutral or even carbon negativity. This not only obliges them to reduce their own emissions, but it also allows them to invest in carbon offsetting projects such as reforestation and renewable energy initiatives to mitigate the environmental impact

Edge data centres support AI

Adopting AI technologies requires a lot of computational power, storage space and low-latency networking to be able to train and run models. These technologies prefer hosting environments, which makes them highly compatible with data centres, therefore, as the demand for AI grows, so will the demand for data centres. However, the challenge remains on limiting new data centres to connect to the grid, which will impact data centre build out.

This highlights edge data centres as the solution to the data centre capacity problem. As it becomes more challenging to build news sites in Frankfurt, London, Amsterdam, Paris and Dublin (FLAP-D), and other key locations, some of the capacity can be moved to edge data centres.

By 2028, its estimated that 50% of AI workloads could be moved to the edge. Edge capacity is predicted to support a number of leading AI use cases, such as Computer vision. Computer vision can particularly benefit from edge data centres due to number of requirements including:

Low latency. The nature of Computer vision applications needs to make rapid decisions based on real-time data from cameras and sensors. For example, in driving assistance systems, they require immediate response to ongoing changes to the likes of road conditions to ensure safety. If that data was processed closer to the source at the edge, it cuts out the need of it going through the centralised data centre before the source, resulting latency being reduced and enables quicker decision-making.

Data security and privacy. The likes of surveillance and facial recognition applications within Computer Vision involve sensitive data being processed locally to ensure privacy and comply with regulations. Utilising edge data centres allows for the data to be processed closer to the source, which minimises the risk of data breaches.

Improves bandwidth efficiency. With raw video data being transmitted in large volumes from multiple cameras, using a centralised data centre, this can cause strain on the network bandwidth which incurs unreasonable costs. Again, with edge data centres allowing the data to be processed locally, relevant information or analysed insights can be retrieved directly to the central server, which reduces the burden on the network, thus optimises bandwidth usage.

Ultimately, real-time processes and intelligent decision-making are continuing to grow across various industries. Therefore, the use of AI is expected to increase with this, which will also demand data centre capabilities that edge can bring.

Edge data centres aid cloud repatriation

In today’s digital economy that continues to grow, many companies are facing pressures to modernise their operations. With this pressure, cloud computing has emerged as a cornerstone for these modernisation efforts, with companies choosing to move their workloads and applications onto the cloud. This shift has brought challenges for companies relating to them managing costs and ensuring data privacy.

As a result, organisations are considering cloud repatriation as a strategic option. Cloud repatriation is essentially the migration of applications, data and workloads from the public cloud environment back to on-premises or a colocated centre infrastructure. This occurs for many reasons, including:

Costs. Many companies are finding the ongoing expense to maintain resources in the cloud outweighing the initial savings or their usage patterns have changed.

Performance concerns. There are certain workloads that require low latency or high performance that cannot be achieved in the cloud space, due to limitations in the network.

Complying to regulations. There are organisations that need to comply with specific regulations that restrict their flexibility with their data, which demands them to migrate from the cloud back to on-premises infrastructure or edge.

These concerns can be addressed with edge data centres. Not only that, they can also maintain agility and scalability offered by cloud computing. With a growing interest in cloud repatriation from certain enterprises, this will help drive demand for edge data centres. 

Mark Turner is Chief Commercial Officer at Pulsant, a provider of UK-based edge colocation centres

Main image courtesy of iStockPhoto.com and ArtemisDiana