Data centre migration - the impact of AI

There are many and various reasons to consider a data centre move: rescaling business, upgrading to a cloud-based or hybrid infrastructure, consolidation after a merger, geographical relocation, tax incentives…the list goes on. One thing is certain though – a data centre migration needs careful planning and execution, so will inevitably consume an organisations time and resources, as well as some level of upheaval. This is not an operation that a business wants to conduct on a regular basis. It’s therefore worth looking at where the data centre industry is heading, and arguably nothing is going to have more impact than the development and deployment of AI technology.

Data centre operations

Essentially, AI and machine learning algorithms excel at recognizing patterns in datasets and applying that learning to future tasks, a process known as predictive analytics. This capability automates and streamlines many day-to-day operations.

Data centre operators are increasingly adopted AI to enhance the efficiency of their daily services. According to a recent survey, 57% of data centre owners expressed trust in AI models for making operational decisions – a nearly 20% rise from the previous year.

Predictive analytics enables real-time enhancements in critical areas such as data centre cooling systems. The substantial heat generated by meeting the power and storage needs of modern computing can be efficiently managed using AI, leading to cost reductions and improved energy efficiency. Notably, Google’s implementation of AI resulted in a massive 40% reduction in cooling costs.

AI also plays an effective role in mitigating IT infrastructure inefficiencies. Predictive analytics aids providers in optimizing power allocation and rack space, contributing to lower operational costs, improved power usage effectiveness, and more data-driven decision-making.

Resource management optimisation

Modern businesses are placing substantial demands on data centre infrastructure. Bearing the  overnight take-up of ChatGPT in mind, a recent report forecasts that the training and implementation of generative AI models will reach a cool $76 billion by 2028. This surpasses twice the annual cost of Amazon Web Services, the world’s largest public cloud provider.

For data centres, this reality necessitates a heightened efficiency across various fronts to meet the power, storage, and connectivity needs of existing, evolving and future technologies. Data centre service providers are already initiating efforts in this direction.

Data centres are also deploying AI for enhanced workload management and allocation, optimizing the utilisation of hardware and network services to prevent downtime and ensure a consistently high level of service.

Predictive maintenance facilitated by AI algorithms plays a crucial role in ensuring business continuity by identifying potential issues before they occur, so significantly reduces downtime and costs associated with hardware replacement. McKinsey estimates that the integration of predictive AI models with appropriate Internet of Things (IoT) devices has enabled certain companies to slash maintenance costs by up to 25%.

Another application of AI is dynamic workload management, which contributes to an improved service by efficiently allocating computational tasks to the most suitable resources. This not only lowers costs but also ensures the best possible customer experience.

Ultimately, AI is instrumental in optimizing data centre infrastructure performance to its maximum potential, benefiting both providers and end-users.

Security enhancements

AI can play a pivotal role in strengthening data centre security, especially in the face of continually evolving cyber threats that may challenge traditional security systems by adapting swiftly or implementing effective defence measures. Employing AI empowers data centres to take a proactive security position.

In particular, data centre AI contributes to enhanced security in the following ways:

• Anomaly detection

AI systems monitoring network traffic, access logs, and system behaviour excel at identifying unusual patterns, enabling real-time detection of potential issues before they escalate. This proactive approach allows security teams to mitigate risks before any significant damage occurs.

• Predictive security measures

Unlike traditional reactive security measures, AI analysis empowers data centres to predict potential threats and vulnerabilities. This capability enables them to close gaps in defences before cyber-attacks can exploit them.

• Data protection

Utilizing AI algorithms and techniques, data centres can optimize data processing, storage, and security. This improvement safeguards the business-critical aspects of uptime, reliability, and data integrity, both during transit and storage.

The future of data centres

AI has already made a significant impact on every industry, and it’s safe to say that this will accelerate at an exponential rate in the coming years. Future data centres must and will adapt in tandem with technological advancements to deliver an efficient and effective service.

Anticipating the next generation of data centres, we might witness the influence of advanced AI, quantum computing, and other emerging technologies. These cutting-edge technologies hold the promise of providing greater efficiency and introducing advanced features.

In 2023, the term ‘AI’ entered the Oxford English Dictionary, and has been the hot topic of the year. As AI becomes integral to data centre operations, concerns related to transparency and accountability will naturally come to the forefront. Sustainability will also play a pivotal role in decisions regarding data centre AI. Data centre energy consumption is projected to increase by 12% by 2030, as workloads with high resource requirements utilise advanced CPUs and GPUs, necessitating sophisticated liquid cooling systems to prevent hardware damage. It’s no surprise, therefore, that AI is already being developed and deployed to manage and create new cooling systems.

AI will also be tasked with the data migration process itself. This is already happening with the major players in the hyperscale data centre sector (see https://www.cio.com/article/654503/salesforce-it-injects-generative-ai-to-ease-its-massive-datacenter-migration.html). It’s simply a matter of time before this trickles down to the SME sector.