Is edge computing gaining popularity?
Increasing digitalisation, coupled with exponential growth in data, has forced companies to rethink data centre resiliency and energy efficiency. What we see in the market is a combination of cloud and edge computing, and this demands low-latency and bringing the computing closer to end-users and devices where the data is produced.
Mega data centres are coming up everywhere. It can be a hyperscaler such as Microsoft or a central data centre of the customer. To a large extent, these are cloud data centres, but in the Middle East, many customers still prefer on-prem or hybrid solutions. We also have regional data centres of large customers, which could be in different countries or cities, like, for example, multi-national banks who must be closer to their customers.
If you look at Formula 1, a lot of data is processed immediately. Now, imagine this data being transferred to a mega data centre for processing. It could be problematic because of bandwidth and latency issues, so we have deployed our modular data centre solutions at many Formula 1 racing locations for faster data analytics. It can be a micro data centre or just a rack server system with a PDU connected as the edge.
This solution can also be used by banks to connect all branches right up to the level of ATMs, by retail shops, or even by global supply chain companies.
Thus, Bank of Lebanon, which had connected all its branches, decided to keep all their data near the data source instead of transferring it to a central data centre for processing and analysis, which helped them to increase response times and reduce bandwidth requirements.
Gartner has estimated that 75 percent of all enterprise data will be processed at the edge by 2025. But you cannot do edge without having a centralised core because you want your IT team managing and monitoring your IT infrastructure from one location. It is also important to remember that the cloud and edge will not replace one another.
Are you focusing on the prefabricated modular data centre market?
We do offer prefabricated data centre systems, right from what we call micro data centres with or without cooling right up to solutions for large data centres. What differentiates us is that we offer software, which provides insights into your data centre environment and gives real-time recommendations to optimise infrastructure performance.
DCIM was used just as a tool. Do you see it evolving as a platform now?
It is already a platform. If you look at our EcoStruxure platform, it is much more than DCIM because we have added analytics, preventive maintenance and digital services to it. What is more, it is a vendor-neutral platform that allows you to manage and monitor your critical infrastructure, whether it is on-prem, in the cloud or at the edge.
Are you leveraging IoT for power distribution and management?
EcoStruxure is an IoT-enabled platform, and we do have solutions for non-IT areas as well. For example, we have a solution that connects petrol stations into one network and manages them centrally. Similarly, we have IoT-enabled edge solutions for mining, oil and gas, airports, and fleet management.
What are the implications of AI on energy-efficiency given the fact these are high-density systems?
It’s a challenge. Data is the foundation of AI, and you need more power to process large datasets, but it is also essential to focus on the energy efficiency of AI systems in data centres. That’s why we ask our customers to replace their server rooms with integrated solutions that feature power, cooling racks and management.
Do you offer liquid cooling systems?
Not at the moment, but we are working with a partner to develop such systems. What we offer, on top of traditional cooling, is free cooling with our Air Economiser system. It requires zero white space and is a rooftop system which maximises the cooling potential of the ambient temperatures outside your data centre.
Discussion about this post