How AI advancements and growth are reshaping the data centre landscape

Author: Beatrice

Three quarters of UK IT decision-makers expect the volume of data they manage to at least double in the next ten years, according to research by Telehouse – and that’s no surprise when considering that data creation, storage and replication is continuing to grow at an unprecedented rate. Many sectors were either completely halted or impacted negatively by the events of the past 15 months but the continued generation of data, cloud migration and technology adoption has pushed data centre growth into uncharted territory, with European data centre capacity set to rise by 21% this year.

Surges in customer demand are creating a need for data to be processed in a more functional and efficient way while effectively managing the increased strain on premises themselves. While a range of supportive technologies are emerging to help achieve this, artificial intelligence (AI) and its numerous possibilities is gaining recognition across the industry. IT decision-makers have even highlighted the technology as one of the top five focus areas for their organisation (27%).

Adapting to automation

While AI has been around for a while in terms of electrical mechanical infrastructure, it can go much further in driving automated data centre operations. Though many data centres have to this point relied on manual operations to power operational decisions, AI can assist in the transfer of load or intelligent switching between redundant and resilient equipment. This allows operational resource within the premises to focus on maintenance and repairs as opposed to the plant running cycles.

With the increased pressure on resources as data volumes rise, the data collected on facility temperatures, humidity levels and the strain placed on power and cooling infrastructure can help data centre operators to be more informed on how to manage the operating life of equipment. They can also calculate where savings could be made in terms of capital expenditure and investment on replacement parts. The result for operators is increased uptime and efficiency, with these benefits then passed on to customers.

Scrutiny around sustainability and security

Improving efficiency with use of AI (and by extension machine learning) naturally extends into what AI can offer in terms of improving data centre sustainability. While arguably data centres have been highlighted for ever increased power consumption and have been under continued scrutiny as a result, many are at the core of harnessing renewable energies while driving internal efficiencies. Managing increased load via AI-powered cooling that can be ramped up and down based on demand is one method of ensuring that energy efficiency gains are made within the data halls. The resulting benefits can both be applied in terms of reducing electricity overheads and helping to bring down overall CO2 emissions.

Also pertinent in the sustainability space is the ability to apply AI to balance equipment workloads and distribute resources accordingly. While running chillers at a reduced capacity helps to keep electricity costs low for data centre operations, it also means that they run inefficiently. AI can identify where data loads are mismatched and modulate output to optimise the whole system, with the resulting data helping to dictate and drive how future cooling plants are designed.

Beyond sustainability, AI-driven solutions, powered by growing amounts of data, are also shaping data centre security measures. AI-powered network management and cyber security can help organisations to better manage any unexpected activity outside of usual traffic patterns, such as large volumes of data being taken from an office network at an unusual time of day. The technology can flag this action and disable the device’s network access to prevent a potential breach of data, while using the incident to learn the typical behaviour patterns of that device to monitor for any future cyber threats.

As data levels grow and energy consumption rises, leading to increased scrutiny from a security perspective, protecting physical server operations in a data centre premises will prove to be just as important as software operations in the future. AI, for example, has the ability to integrate with network monitoring to check for any changes in server behaviour even after a person has entered a secure server hall.

Future potential

Data-driven AI advancements are no doubt already benefitting the data centre sector in numerous ways, and facilitating an exciting evolution across the entire landscape. Even more exciting, however, is that it will continue to do so in the coming years with the advancement of predictive modelling, of which the sector is currently only scratching the surface. Predictive modelling will allow operators to make beneficial changes to its data centre estates well in advance, such as identifying and facilitating an extra 1 or 2% of operational efficiency within the plant. While seemingly small, this will lead to significant cost savings and operational advantages that can benefit both operators and their customers. These advancements are only becoming possible now due to the past efforts of the data centre sector to measure and record vast data sets (collectively referred to as data lakes), ready for analysis and intelligent use at a later date.

Data may be growing at an exponential rate, but data centres, while facing pressure due to the increased network strain, are also benefitting from its incorporation into AI technologies to boost automation, drive sustainability and enhance security measures for the industry. The next step for AI is to play a primary role in driving better decision-making, bringing efficiencies that were previously not possible and empowering the sector as a whole.

By Oliver Goodman, Head of Engineering at Telehouse