Energy efficiency and sustainability in telecommunications: How CSPs can leverage DPI to cut costs and up profits

Energy efficiency is a global priority. A sense of urgency now looms across various industries, including telecommunications, concerning the consequences of unchecked energy usage and wastage. According to the GSMA, communication service providers (CSPs) consumed between 2% to 3% of total global energy in 2021¹, and in 2020, the telecoms industry was reported to contribute to 2.6% of global carbon dioxide emissions². Therefore, improving energy efficiency in telecommunications is a growing concern for CSPs worldwide.

Networks form a major part of the telecom business. The Radio Access Network (RAN), for example, accounts for 72% of total CSP energy consumption³. Thus, improving network energy efficiency has a significant effect on the industry’s overall energy consumption. Energy use in CSP networks comes from transmission and computing, which form the core, edge and access networks. Computing, in particular, is power- and water-intensive. A mid-sized data center uses 300,000 gallons of water a day, equal to the water consumption of 100,000 homes⁴, according to ING.

Energy use in telecoms, unfortunately, is not slowing down anytime soon. In 2024, mobile networks carried over 140 EB of data monthly⁵. Densification of 5G networks is expected to exacerbate this issue given the forecasted five-fold growth in traffic and the fact that 5G base stations need at least twice the power of 4G⁶. The adoption of AI by CSPs is also pushing energy usage in the industry. Goldman Sachs Research estimates that by 2028, AI will account for 19% of data center power usage⁷.

The balancing game

In reality, short-term profitability and user experience concerns often overshadow CSP energy efficiency goals as well as their compliance with regulations, such as EU’s rating scheme for data centers and GSMA’s 2025 net zero emissions target. This is where operators see the increasing importance of having in place intelligent networks where data-driven optimization allows energy consumption to be regulated and minimized. Data-driven optimization in turn, relies on real-time traffic visibility, which helps CSPs understand how traffic flows influence energy usage and costs.

Leveraging such insights, CSPs can implement a ‘divide and conquer’ strategy which accords different policies to different traffic types. It allows CSPs to turn on application-based prioritization that ensures premium pathways are reserved only for latency-sensitive applications and high-SLA customers. CSPs can also compress content-heavy applications such as video streaming and bring down transmission loads. Similarly, they can cache popular content and cut down content fetches.

Network optimization through DPI

Traffic visibility also speeds up traffic processing as CSPs become immediately aware of what they are handling. By cutting down the interludes between one flow to another, thanks to real-time identification of applications, traffic visibility allows clearance of traffic streams, which saves on network resources in use, at any one time. Thus, by leveraging DPI technology CSPs can optimize their network resource allocation, reduce energy consumption, and improve overall network efficiency.

Traffic intelligence additionally allows CSPs to pre-filter traffic. This enables their data pipes to drop packets that are redundant, malware-infested, anomalous or that involve forbidden tethering or rogue devices. Traffic filtering in RAN, for instance, immediately reduces the network load on edge and core networks.

Optimizing network resource allocation via traffic analytics

Equipped with real-time traffic intelligence, CSPs can compute and observe performance, resource usage and security metrics across the network. These insights can uncover poor architectures, tackle loss of resources due to attacks and abuse, and remove duplicate processes that are often hidden in fragmented systems. By fixing these gaps, CSPs can trim their traffic flows and eliminate energy wastage. These insights also enable CSPs to continuously adapt their network architectures based on various efficiency metrics.

Deep packet inspection for telecom sustainability

Deep packet inspection (DPI) is a traffic detection technology that enables CSPs to gain real-time traffic visibility and optimize their networks for improved energy efficiency. Its granular and accurate insights form the basis of many network policies and decisions. Lightweight and high-performance DPI engines, such as ipoque’s R&S®PACE 2 and R&S®vPACE, offer protocol, application and service-level insights that enable CSPs to monitor and optimize energy usage across the network. These engines also represent a lean, energy-efficient implementation using the following specifications:

• High speed: R&S®PACE 2 can run multiple double-digit Gbps per core, ensuring low latency processing even in demanding network environments.
• Low memory consumption: Both engines boast the lowest memory footprint in the market and feature individual optimization options.
• Efficient thread and process management: With support for multi-threading and uniform (UMA) and non-uniform memory access (NUMA), both engines support nearly linear scalability to the number of CPUs available.
• First packet classification: Via service and DNS caching, our DPI engines classify an IP flow from the very first packet, enabling subsequent packets to bypass the filtering process entirely.
• Support for vector packet processing (VPP): Optimized for cloud-native VPP frameworks (e.g. FD.io, DPDK Graph and RF_Ping), R&S®vPACE enables 3x higher performance than scalar packet processing DPI engines, as well as a reduced clocks-per-packet count.
• Architecture-agnostic: Our DPI technology is deployable on any Unix-based operating system and on any hardware with a C compiler in a cloud, virtualized or traditional network environment. This removes the need for additional devices or infrastructure.
• Comprehensive coverage: By leveraging encrypted traffic intelligence (ETI), which is based on machine leaning and deep learning, both engines enable CSPs to filter every flow, including encrypted, obfuscated and anonymized flows.

Promoting energy efficiency in network slicing and video traffic management for improved sustainability

CSPs can leverage ipoque’s cutting-edge DPI technology to make a huge difference in their sustainability outcomes. Let’s take the case of network slicing: the instantiation of virtual slices requires the orchestration of different functionalities that are mapped to each different application. R&S®PACE 2 and R&S®vPACE enable CSPs to build and configure slices quickly, and map different applications to different network functions and network resources, at scale. This cuts down processing times and minimizes the consumption of power and computing. By adopting sustainable network practices for energy-efficient network slicing and video traffic management, CSPs can reduce their environmental impact while improving their bottom line.

Another use case is the transmission of video traffic. IoT applications such as CCTV cameras use up network resources rapidly as bandwidth-intensive content is transported all the way from access networks to the core. DPI engines, such as R&S®PACE 2 and R&S®vPACE, can rapidly identify video packets from selected applications for edge pre-processing or processing at the source. ipoque’s DPI technology not only helps in reducing the load on subsequent nodes, its form factor and lightweight quality keeps processing overheads in edge nodes to a minimum.

Promoting energy efficiency for energy cost reduction in telecommunications

As the telecommunications industry focuses on reducing energy costs, traffic visibility will soon become an integral part of managing CSP networks. Highly efficient traffic visibility tools, such as R&S®PACE 2 and R&S®vPACE, provide CSPs the insights needed to realize incremental cost and resource savings over time, enabling them to gradually reduce the network total cost of ownership (TCO) and enhance their long-term profitability and sustainability.

Sources:

[1] GSMA Intelligence: "Going green: energy efficiency in telecoms"
[2] TM Forum Inform: "Telco cloud energy optimization for a sustainable future"
[3] MWC GSMA: Planning is Crucial for the Sustainability of the RAN
[4] ING: "Growth in water consumption of data centres needs more attention"
[5] Ericsson: Mobile network traffic Q1 2024
[6] VIAVI: "What is 5G Energy Consumption?"
[7] Goldman Sachs: "AI is poised to drive 160% increase in data center power demand"