5G is shaking up the data center world

Increased speeds (a few seconds to download 1GB of data), reduced latency (1 ms compared to 25 to 40 ms for 4G, thanks to high-density servers for data processing and analysis), and improved user experience in mobile situations: 5G offers a glimpse of new uses and business opportunities that were previously technically impossible in all sectors of activity. The corollary is an exponential increase in global mobile data traffic. Indeed, according to a study by Cisco (Global Mobile Data Traffic Forecast Update), the latter is expected to increase sevenfold and grow almost twice as fast as fixed IP traffic between 2017 and 2022.

This considerable increase in data volume can be explained by a shift in usage in recent years: the development of streaming, high-resolution video viewing (thanks to full HD screens, 4K in the short term, and 8K in the future), virtual reality, remote-controlled machines in real time (Industrial Internet of Things, M2M), and connected artificial intelligence (home automation, smart buildings, etc.). Meanwhile, the breakthrough of Cloud gaming, which is very demanding in terms of latency, is also expected to see strong acceleration. Not to mention new uses, still unknown to this day...

This announced (r)evolution involves reviewing strategies for managing an ever-increasing volume of data. Today, central data centers support the bulk of the data volume generated by the mobile internet. However, with the increase in traffic, operators' backbone networks will inevitably become saturated.

Operators will be able to avoid this saturation of networks and data centers provided they quickly adopt new strategies for sizing and integrating new technologies adapted to these developments. This notably involves bringing data closer to the antennas to guarantee satisfactory latency.

This new approach involves increasing the number of local data centers (in the regions) and deploying technologies dedicated to Edge computing. The creation of these decentralized mini data centers (or Edge data centers), linked to existing infrastructure, will offer a better distribution of storage and computing locations, so as to relieve national networks.

Furthermore, the creation of such a network of data centers, combined with the acceleration of the 5G and IoT industry, should contribute to the economic development of territories outside major cities and, consequently, to the creation of jobs resulting from the associated operation and maintenance needs.

However, inside the data centers themselves, a certain continuity will be required regarding their management: whether micro or "classic," their urbanization will remain essential to optimize energy consumption, take full advantage of their surface area, and anticipate future developments.

In other words, the data center world is now adapting to the requirements of operators and evolving according to the autonomy needs of industrial players. In this context, and in order to fully exploit the potential of 5G, it is necessary to build a real strategy to be able to process and store more data at the edge securely. All this must be done within an energy-efficiency framework, which is all the more essential as the impact of the digital sector is growing and, in this respect, involves the responsibility of all stakeholders: operators, content providers, infrastructure builders, terminal manufacturers, and consumers.

By Ali Alaeldine, Project Director, APL.

Ali Alaeldine is a Project Director at APL, holds a PhD in telecommunications systems from INSA Rennes, and is a corporate strategy expert graduated from HEC Paris. He has 18 years of experience in project direction and management and team management, including eight years in data center environments. He has participated in strategic projects at national and international levels for companies such as SFR, Automic, Thales, Airbus, and Stanley Black & Decker. Ali Alaeldine is a certified trainer in cloud computing, SAM, and ITIL, and is certified in Lean IT, AgilePM, and as a PRINCE2 Practitioner. He holds an engineering degree in information systems and networks and a research master's degree in instrumentation and microelectronics from the University of Metz.