Clik here to view.
Open radio access network (O-RAN) technology is driving the mobile communication industry toward an open, virtualized, and disaggregated architecture. O-RAN breaks traditional hardware-centric RANs into building blocks—radios, hardware, and virtualized functions—enabling mobile network operators to create their RANs using a multivendor, interoperable, and autonomous supply ecosystem. Using open and standardized interfaces, O-RAN enables network vendors to focus on specific building blocks rather than creating an entire RAN. Similarly, operators can mix and match components from multiple vendors. While O-RAN delivers many improvements, energy efficiency is a top priority.
Energy efficiency in RAN
Global efforts toward a carbon-neutral future and consumer demand for greener products have increased the urgency to focus on energy efficiency. Sustainability is a critical priority for the information and communications technology (ICT) industry, which is committed to making 6G and 5G a green reality.
Three key performance indicators define objectives and characterize improvements for a RAN energy optimization effort:
- Energy consumption (EC) represents the energy used to power the infrastructure. The European Telecommunications Standards Institute (ETSI) ES 202 706-1 defines EC as the integral of power consumption.
- Energy savings (ES) represents the reduction of energy consumed with minimal impact on the quality of service (QoS).
- Energy efficiency (EE) refers, in a general sense, to a measure of how an appliance or system uses energy. EE is the ratio between useful output or service over the required energy input.
These three factors are essential to consider. EE improvement strategies aim to apply several mechanisms when there is no need for all the available performance, thereby minimizing the impact on QoS and the user experience. Engineers need a balanced approach depending on QoS goals. They must make insightful measurements to understand power consumption rates for different load conditions and metrics.
Energy efficiency in O-RAN
The ETSI ES 203 228 test specification considers the gNodeB as a whole. However, the O-RAN Alliance® recognizes the urgency of addressing EE and EC in a disaggregated RAN. For example, the initial version of the O-RAN fronthaul interface specification included signaling mechanisms to notify the radio unit about periods of non-usage of radio symbols. These signaling mechanisms enabled the radio to halt transmission and conserve power. The fronthaul interface now provides the capability to inform the network about energy-saving capabilities in each radio, such as carrier deactivation, enabling automated activation and deactivation of energy-saving mechanisms. The O-RAN Alliance is also developing energy-saving test cases to ensure conformance and enhance vendor interoperability.
As shown in Figure 1, energy consumption spans the entire network, including the grid, RAN, core, and transport, depending on many parameters: from RF channels to topology. Therefore, energy consumption requires a comprehensive approach involving multiple parts of an operator’s organization to capture all components. Test engineers must consider and tweak numerous parameters and variables to identify optimal configurations. The main question remains: How can test engineers reduce energy consumption and costs without impacting QoS?
Image may be NSFW.
Clik here to view.
Figure 1 O-RAN architecture with user equipment (UE) and core network. Source: Keysight
Reducing energy consumption and cost without impacting QoS
1. Reducing energy consumption
There are numerous techniques to reduce EC at a network level, each requiring varying levels of effort for implementation. Migrating technology from legacy platforms onto the most recent and energy-efficient platforms can immediately reduce network energy consumption. Such a migration requires an upfront investment in new equipment and resources to perform the upgrade, but it is a relatively low engineering effort. If investing in equipment upgrades is not possible, analyzing existing deployments, eliminating redundancies, and identifying overprovisioned devices helps improve energy consumption.
While it requires a mix of engineering and equipment investments, network topology optimization can also help reduce EC by determining the ideal minimum subset of equipment necessary to cover different topologies without sacrificing the QoS.
2. Optimizing energy efficiency
To maximize energy efficiency, engineers must continuously adapt user demand to the supply of network resources. They need to dynamically allocate the correct number of computing services and radio resources to match demand and aggregate user demand to ensure the entire use of each resource. For example, engineers can avoid using two servers at 50% load each. By applying the methodology at various levels, from the system to the device/chipset levels, engineers can optimize EE.
System-level intelligence is another way to maximize EE by performing dynamic resource allocation decisions at the system level. Engineers would activate or turn off nodes on wireless networks and perform load balancing to redirect users to active nodes. Similarly, they can allocate resources to the device hardware and chipset level.
Semiconductors and chipsets are the first elements of the energy chain. Hence, they are the main contributors to energy consumption and efficiency. New chipset generations provide advanced resource optimization capabilities, such as turning on or off discrete digital resources on the chip. Engineers can accomplish this mechanism by changing analog parameters (clock speed and bandwidth) to adjust the desired performance level and reduce power consumption since the energy is a function of electrical transitions in each gate.
At the radio level, engineers can perform additional optimization with innovative scheduling capabilities in the O-RAN distributed unit when traffic is low. They can regroup physical resource blocks (PRBs) from multiple symbols into a reduced number and augment the transmission blanking time.
Optimizing EE requires using chipset-level power-saving capabilities to the fullest extent possible. Only then can engineers determine the hosting of the power decision entity to prevent conflicts.
3. Standardization
The emergence of standardized O-RAN drives the need to define standards that enable intelligent control and energy optimization of multivendor-based networks. Ultimately, the Alliance’s work will result in new O-RAN specifications and technical reports sections. The Alliance’s ongoing work includes defining procedures, methodology, use cases, and test case definitions for cell/carrier switch on/off, RF channel selection, advanced sleep modes, and cloud resource management.
4. Embedded and chipset-based energy optimization
Intelligent control loops significantly contribute to energy optimization at the system level. But these loops are also appropriate at the chipset level as they contribute to local power optimization within a device.
Chipset sleep mode mechanisms consist of deactivating or slowing down function within a specified period. Different sleep levels enable multiple levels of energy saving.
However, each sleep mode comes at a cost: the deeper the sleep and energy-saving, the more time the chipset remains in the sleep mode and the wake-up transitions. These transitions are not energy-efficient and may offset gains from the sleep phase. Therefore, defining sleep strategies optimizes the trade-off between transitions and sleep phases.
How to Measure and evaluate the energy efficiency of O-RAN components
Figure 2 shows an O-RAN architecture which consists of the following components:
- O-RAN radio unit (O-RU) for processing the lower part of the physical layer
- O-RAN distributed unit (O-DU) for baseband processing, scheduling, radio link control, medium access control, and the upper part of the physical layer
- O-RAN central unit (O-CU) for the packet data convergence protocol layer
- O-RAN intelligent controller to gather information from the network and perform the necessary optimization tasks
Image may be NSFW.
Clik here to view.
Figure 2 An overview of O-RAN architecture with the O-RU for processing the lower part of the PHY layer, O-DU for processing the upper part of the PHY layer, O-CU for the packet data convergence protocol layer, and an O-RAN intelligent controller to perform optimization. Source: Keysight
Energy plane testing requires a cross-domain measurement system and cross-correlation of the data to gain meaningful insights into the energy performance of the RAN components. The testing combines power measurement with protocol and RF domains. As O-RAN and the 3rd Generation Partnership Project (3GPP) are fast-evolving standards, equipment manufacturers must ensure product compliance with the latest versions. Automation of the test cases and report generation are the keys to ensuring compatibility with the latest standards of regression testing.
Measure the energy efficiency of an O-RU
RAN energy consumption and efficiency improvement requires minimizing power usage while maximizing performance. For RU testing, the ETSI ES 202 706 standard, which describes the test methodology to measure power consumption in a gNodeB, can be adapted to make similar measurements in an O-RU under different load conditions, representing a typical day in the life of a RU—the load changes during the test in low, medium, high, and complete steps (Figure 3). So, by measuring the O-RU at different loads, we can calculate the total energy consumed.
Image may be NSFW.
Clik here to view.
Figure 3 Decoded constellation, signal spectrum, allocated PRBs, EVM per modulation type and decoded bits. Source: Keysight
To measure the energy efficiency of an O-RU, test engineers need an O-DU emulator, a DC power supply, and an RF power sensor (Figure 4). The O-DU emulator generates different static traffic levels from low, medium, busy, to full load traffic as defined by the ETSI ES 202 706-1 standard. A DC power supply provides power to the O-RU and measures the accumulated power consumption over time. The RF power sensor measures the output power at the antenna connector port. The ratio of output RF power to input DC power represents the energy efficiency measurement.
Image may be NSFW.
Clik here to view.
Figure 4 O-RU test set-up with an O-DU emulator, power sensor, as well as a power supply and analyzer. Source: Keysight
Measure the energy efficiency of O-CU/O-DU
Ensuring accurate and standardized EE of O-DU and O-CU in an O-RAN involves assessing various factors related to power consumption, resource utilization, and overall network performance. As EE is the ratio of delivered bits and consumed energy, test engineers need access to the user equipment (UE) throughput data to ensure that lower EC is not at the cost of lower quality of service. The fronthaul and backhaul interface require emulation to measure the EE of an O-DU and O-CU. In addition, test engineers must be able to simulate the traffic profiles of different pieces of UE.
The fronthaul requires an O-RU emulator to provide the interface to the O-DU. A UE emulator simulates the traffic flow to the O-RU emulator the UEs request. The backhaul requires a core emulator or a live core network. An AC or DC power supply capable of recording the output power measures the combined energy consumption of O-DU / O-CU. The ETSI specification does not refer to the disaggregated base station architecture, so the points of power measurement can vary depending on the implementation. The test software generates an energy efficiency report by simulating different UE traffic profiles with varying path loss, file size, and throughput.
Evaluate the performance of gNodeB
To test gNodeB, test engineers can use a set of automated test cases and analytics tools based on ETSI standards. The test setup should include a UE emulator, a core network emulator, and a power analyzer. The UE emulator emulates stateful UE traffic and measurements, while the core network emulator terminates the calls from the UE emulator for stateful O-DU/O-CU testing. Both emulators require dimensioning to load testing scenarios, and a power analyzer measures the server’s power consumption.
Energy efficient wireless networks
While the wireless communication industry increasingly prioritizes sustainability and net zero strategies, achieving energy efficiency has become as important as performance, reliability, and security. As wireless networks evolve into multivendor disaggregated systems, collaboration among chipsets, equipment, and test vendors is necessary to optimize power consumption without compromising performance.
Moving forward, test and measurement companies should focus on delivering cutting-edge technology and tools that accelerate the transition to green and sustainable wireless communications, realizing the network performance and capital expenditure (CapEx) advantages of O-RAN. To achieve that, understanding the energy performance of RAN components is key. As highlighted in this article, there are methodologies providing standardized and accurate assessments of energy efficiency in RAN components, essential for optimizing network performance while minimizing energy consumption in the increasingly dynamic and complex telecommunications landscape.
Image may be NSFW.
Clik here to view.
Chaimaa Aarab is a use case marketer focused on the wireless industry (5G, 6G, Wi-Fi 7, O-RAN) at Keysight Technologies. Her background is in electronics engineering with previous experience as a technical support engineer and market industry manager.
Related Content
- Creating more energy-efficient mobile networks with O-RAN
- O-RAN is transforming 5G network design and component interoperability
- O-RAN challenges from the fronthaul
- Making waves: Engineering a spectrum revolution for 6G
- 5G NR base stations bring new conformance testing challenges
The post Navigating energy efficiency in O-RAN architectures appeared first on EDN.