Design and Analysis of Power Consumption Models for Open-RAN Architectures

Urooj Tariq, Rishu Raj, and Dan Kilper · 2025

Read Design and Analysis of Power Consumption Models for Open-RAN Architectures by Urooj Tariq, Rishu Raj, and Dan Kilper (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-24552.txt.

Introduction

Urooj Tariq, Rishu Raj, and Dan Kilper model power use in open radio access networks in this 2025 study. The O-RAN design splits a mobile network into parts that can come from different makers and be run from shared data centers. The authors ask how much power the full path uses, from the radio unit at the antenna to the data center.

Their models count power step by step along that path and test where the heavy baseband processing should sit. Tight limits on how many radios one unit can serve push processing toward the center of the network. Letting each unit serve more radios cuts that need, but the links then carry far more data and use more energy for transport. Network planners can use this trade-off to weigh energy costs when they lay out new systems.

Catalog record

CREATORUrooj Tariq, Rishu Raj, and Dan Kilper
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE27 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-24552

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Design and Analysis of Power Consumption Models for Open-RAN Architectures Urooj Tariq, Rishu Raj, and Dan Kilper Abstract The open radio access network (O-RAN) Alliance developed an architecture and specifications for open and disaggregated cellular networks including many elements that are being widely adopted and implemented in both commercial and research networks. In this paper, we develop transaction-based power consumption models of a centralized O-RAN architecture based on commercial hardware and considering the full end-to-end data path from the radio unit to the data center. We focus on recent fanout limitations and early baseband processing requirements related to current implementations of O-RAN and assess the power consumption impact when baseband processing is employed at different centralization points in the network. Additionally, we explore how greater fanout and sharing deeper into the network impact the balance of processing and transmission. Low processing fanout restrictions motivate greater centralization of the processing. At the same time, allowing for more open radio units per open distributed unit will quickly increase the transmission capacity requirements and related energy use. CONNECT Centre, Trinity College Dublin, Ireland; Email: {tariqu, rajr, dan.kilper}@tcd.ie I Introduction Communication networks are evolving in response to emerging connectivity paradigms such as artificial intelligence (AI) native networks that provide greater functionality and reach. This evolution has been accompanied by important architectural changes allowing for greater centralization and virtualization of network functions. A greater reliance on virtualization has also enabled open networking standards such as the open radio access network (O-RAN) specification [ 1 ] . O-RAN allows for the use of network control and management applications that could be offered by third party vendors. It leverages open standards and interfaces to decouple hardware and software components. This enables network operators to mix and match solutions from different vendors, increasing flexibility and reducing dependency on a single supplier. The potential benefits include cost reduction, faster innovation through competition, easier scalability, and a more dynamic ecosystem that fosters collaboration. Investigating the power consumption in O-RANs is essential for optimizing energy efficiency, reducing operational costs, and minimizing environmental impact, thereby making 5G/6G and future networks more sustainable and scalable. The O-RAN Alliance has identified energy-efficient network operation as a key early use case for O-RAN network management applications [ 2 ] . Additionally, efficient power management is crucial for meeting green network initiatives and reducing carbon footprints in modern telecommunications. As such many recent studies of energy use in O-RAN based networks have emerged. However, much of this work has focused on the radio access network portion of the network with little or no consideration for the overall end-to-end network energy use, from the radio units to the data center [ 3 ] . The work in [ 4 ] proposes a mixed integer linear programming problem for energy-efficient placement of baseband units in wavelength division multiplexed aggregation networks. Furthermore, optimization and control studies [ 5 ] consider a particular model or objective functions specific to the problem formulation without addressing the details of the energy use of networks and their underlying equipment in operational configurations. Understanding energy use in a deployed network context is particularly important to determine how the equipment configurations and architectures impact the overall network energy use and to estimate the power consumption of current operational networks. Recent O-RAN implementations have been limited to just a few open radio units (O-RUs) fanning out from each open distributed unit (O-DU). Furthermore, the compute resources for baseband processing (BBP) can involve multiple cores per 10 Gb/s O-RU connection [ 6 ] . These factors can have a large impact on the energy efficiency of the overall network including transport and compute. The most commonly used methods for network energy use modeling are the so called ‘top-down’ and ‘bottom-up’ methods. Such analytical models were recently used to analyze and optimize the power consumption of network elements in a fog cloud architecture for 6G networks [ 7 ] . Yet another approach is the transaction-based modeling technique which models data traffic flow through a network based on mean quantities such as the mean number of hops and typical equipment configurations. This type of modeling has been employed to evaluate the energy consumed by a particular service [ 8 ] and to observe the effect of variations in the traffic volume of the network [ 9 ] . Transaction-based models have also been used to calculate optical IP network power [ 10 ] and different access networks in wired and wireless systems [ 11 ] . In this work, we develop transaction-based power consumption models for recent architectures of O-RANs. We formulate the processing power and transmission power individually, then aggregate them to determine the total power consumption when BBP is carried out at various nodes within the network. For 5G, evolved common public radio interface (eCPRI) is used, which includes the digitized radio signals along with control and management data, synchronization signals, and other auxiliary information [ 12 ] . We also analyze how changes in coverage and network topology impact the power consumption of the O-RAN network through such models. This comprehensive analysis provides critical information that aids in the design and optimization of O-RAN architectures. II O-RAN Architecture and Power Consumption Model In this section, we first describe the O-RAN architecture and then present the proposed transaction-based model for the calculation of power consumption in O-RAN networks. II-A Network Architecture O-RAN specifies standardized open interfaces and constructs within a radio access network (RAN) with the aim of facilitating co-existence among ecosystem participants and creating possibilities for new entrants [ 1 ] . The O-RAN architecture provides a highly flexible multi-vendor network by interface standardization and disaggregation of the hardware and software. Fig. 1 illustrates an overview of the C-RAN architecture commonly used with O-RAN divided into three levels, namely, access network, metro network and longhaul network. The access network represents the user edge side of the network consisting of O-RUs, and connects the users to the next level of hierarchy through the fronthaul link. In the metro network, the traffic from a cluster of O-RUs is aggregated through the transport equipment in the fronthaul links that connect O-RUs and O-DUs, as depicted in Fig. 1 . The open central unit (O-CU) is the next node in the hierarchy of the O-RAN network and it gathers traffic from a set of connected O-DUs and communicates with the core of the network through the high capacity core routers. O-DUs connect to the O-CUs via midhaul links or metro backhaul links if BBP has been fully implemented prior to the O-CU. The longhaul network connects large, often warehouse-scale, DCs to perform computationally intensive tasks and also benefits from the highest efficiency computing resource sharing. The high capacity edge router aggregates the traffic from the network through WDM fiber links and connects the metro network to the core network through backhaul links. WDM links are also used to interconnect core nodes through the gateway router and perform routing for the complete network, as illustrated in Fig. 1 . II-B Proposed Power Consumption Model We now describe our model to investigate the power consumption in O-RAN architectures using a transaction-based approach [ 3 ] . In this approach, we estimate

Read the complete text