Telecom networks consume 3% of the globe’s annual power and are responsible for 2% of worldwide greenhouse gas (GHG) emissions. While the thought of operating more sustainably is one thing, achieving environmental, social and governance (ESG) goals is a complex and layered problem. The communication ecosystem will need to join forces to meaningfully tackle its collective environmental impact, according to Red Hat’s Chief Telco Architect Rimma Iontel.

A majority of sustainability-focused projects are deployed in isolation and focus on individual networks or issues. This strategy, however, “has shown its limitations,” Iontel said.

And artificial intelligence (AI) isn’t yet mature enough to solve the efficiency issues plaguing the telecom industry.

Data analytics powered by AI have emerged as one answer to better outcomes for communications service providers. But AI is only as helpful as the data it’s fed, and collecting accurate data on traffic patterns and energy use across traditional, cloud and multi-vendor networks is not yet the norm. It’s “a challenge still to be solved,” Iontel said.

Why AI isn't the answer

Transforming AI recommendations for a network and its associated workflows into action, even with massive amounts of automation, is another lingering challenge. What’s more – AI technology itself can contribute substantially to power consumption. “We need to be very deliberate in its use for network power optimization,” Iontel warned.

Instead, a unified, holistic approach based on open platforms is more fruitful – and Red Hat put that philosophy into practice at Mobile World Congress (MWC) Barcelona 2024.

Open 5G core infrastructure lowers power consumption

At MWC, Red Hat, NEC and Intel announced a partnership aimed at providing an open 5G core infrastructure that lowers the operational costs related to energy rates and power consumption.

NEC is able to reduce the commercial power use of its converged 5G user plan function on Red Hat OpenShift by more than 30% thanks to Intel’s infrastructure power manager for 5G core. With Intel’s contribution, each processor core has access to the optimal amount of power at all times, which lowers emissions and energy bills.

Tackling network energy use with Intel, Ericsson

Red Hat also shared its work with Intel and Ericsson to develop, integrate and deploy sustainable cloud-native technologies that lower network energy consumption and carbon emissions.

With a goal to strike a balance between optimizing and reducing power use while supporting network performance, the vendors initially focused on radio access networks (RAN) and demonstrated 20% reduction in processing power consumption using Intel’s 4th Gen Xeon scalable processors with virtual RAN (vRAN) boost.

The three pillars of this collaboration are hardware, software and automation. Its energy efficient servers and accelerators, dynamic CPU core scaling for the RAN distributed unit and centralized unit workloads and automation based on network utilization and traffic patterns are designed to help service providers realize the benefits of sustainability.

Sustainably scaling telco data centers

Red Hat also partnered with Intracom Telecom to build a solution that optimizes on-premises and cloud-based OpenShift infrastructures by “proactively adjusting server workloads and strategically powering off underutilized servers to address idle power consumption, without compromising workload performance and stability,” Iontel said.

This solution dynamically scales the number of worker nodes based on real-time predictions of resource demand. This approach ensures optimal scalability and efficiency of infrastructure.

The target market for this service is telco data centers that host 5G network functions and other related services. Traffic load for services like user plan function pods, AI pods and edge pods varies cyclically based on daily and weekly patterns.

Red Hat’s smart scaling capabilities, however, have the potential to reduce the average number of worker nodes – particularly in telco data centers with daily load fluctuations – which translates into power and cost savings, according to Iontel.