CollabORAN: A Collaborative rApp–xApp–dApp Control Architecture for Fairness-Adaptive Resource Sharing in O-RAN

Anastasios E. Giannopoulos, Sotirios T. Spantideas, Panagiotis Trakadas · 2026

Read CollabORAN: A Collaborative rApp–xApp–dApp Control Architecture for Fairness-Adaptive Resource Sharing in O-RAN by Anastasios E. Giannopoulos, Sotirios T. Spantideas, Panagiotis Trakadas (2026) 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-2603-20805.txt.

Introduction

Giannopoulos and co-authors present CollabORAN, a layered control design for open radio access networks. Control loops in O-RAN often run alone at one timescale each. Their design joins three app levels in one nested loop. The rApp sets traffic aware policy over minutes. The xApp assigns spectrum with a hypergraph coloring method each second. The dApp at the distributed unit enforces fairness in milliseconds. Simulations showed better fairness and less user starvation. Spectrum reuse stayed efficient in dense, shifting settings.

Catalog record

CREATORAnastasios E. Giannopoulos, Sotirios T. Spantideas, Panagiotis Trakadas
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE38544
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.20805

Browse the full library · About the Tower of Babel

Full text

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

CollabORAN: A Collaborative rApp–xApp–dApp Control Architecture for Fairness-Adaptive Resource Sharing in O-RAN Anastasios E. Giannopoulos, Sotirios T. Spantideas, Panagiotis Trakadas Abstract The evolution of Open Radio Access Networks (O-RAN) enables programmable and intelligent control of radio resources through disaggregated architectures and open interfaces. However, existing solutions typically rely on isolated control loops and fail to jointly address end-to-end optimization objectives across multiple timescales. Thus, it remains a key challenge to functionally split optimization algorithms across timescale-specific O-RAN layers while complying with control loop latency specifications. This article proposes CollabORAN, a collaborative rApp–xApp–dApp hierarchical framework for dynamic and equitable spectrum sharing in O-RAN systems. CollabORAN leverages a nested control structure in which the rApp performs traffic-aware policy generation, the xApp executes interference-aware spectrum allocation via hypergraph-based PRB coloring, and the DU-level dApp enforces temporal fairness through fast scheduling. The proposed end-to-end closed-loop design enables coordinated optimization across minutes, seconds, and millisecond time scales. Simulation results demonstrate that CollabORAN significantly improves service fairness and reduces user starvation while maintaining efficient spectrum reuse in dense and dynamic network environments. I Introduction I-A Modern Open Radio Access Networks The rapid evolution of Open Radio Access Networks (O-RAN) is fundamentally reshaping the way radio resources are managed in modern cellular systems [ 1 ] . By disaggregating the traditional base station architecture and exposing standardized interfaces, O-RAN enables the deployment of intelligent applications that dynamically optimize network operation. Central to this vision is the RAN Intelligent Controller (RIC), which supports programmable control logic through rApps in the Non-Real-Time RIC (Non-RT RIC) and xApps in the Near-Real-Time RIC (Near-RT RIC). These applications allow operators and developers to introduce policy-driven automation and data-driven optimization into the RAN [ 10 ] . Recent advances in O-RAN architectures are further extending this paradigm with the introduction of distributed applications (dApps) operating closer to the radio access nodes, typically within distributed units (DUs) or Central Units (CUs) [ 4 ] . While rApps focus on long-term policy definition (seconds-to-hours timescale) and learning-based optimization, and xApps perform near-real-time control (below 1 sec) across multiple Radio Units (RUs), dApps enable ultra-fast decision making (below 10 ms) directly at the RAN execution layer [ 14 ] . This emerging three-tier intelligence hierarchy (rApp, xApp, and dApp) creates new opportunities for coordinated multi-timescale optimization of radio resources, allowing policies, interference management, and scheduling decisions to be jointly orchestrated across different control layers [ 4 ] . At the same time, cellular networks are facing increasing pressure on the radio spectrum due to the rapid growth of connected devices, heterogeneous deployments, and data-intensive services. Dense multi-cell environments introduce complex interference relationships among users and base stations, which significantly complicates spectrum allocation decisions. Efficient spectrum access mechanisms must therefore address several competing objectives simultaneously, including interference mitigation, efficient spectrum reuse, and fair resource distribution among users [ 2 ] . I-B Gaps, Challenges and Contributions Existing resource allocation approaches in O-RAN environments typically rely on either static scheduling policies implemented at the base station level or isolated optimization mechanisms implemented within individual xApps [ 2 , 15 ] . Static schedulers are inherently limited in their ability to adapt to dynamic traffic demand, user mobility and interference conditions in dense and heterogeneous deployments. At the same time, most current O-RAN control solutions focus on single-loop optimization performed at the Near-RT RIC, where an xApp attempts to optimize specific network objectives such as interference mitigation, traffic steering, load balancing, or throughput maximization [ 3 ] . While xApps enable near-real-time control across multiple cells, relying solely on xApp-level optimization introduces several limitations. First, xApps typically operate on short time horizons and limited network context, which restricts their ability to incorporate long-term policy objectives such as service-level agreements (SLAs), traffic forecasts, or operator-defined fairness targets. Second, xApps generally assume that resource scheduling and enforcement occur locally at the RAN nodes, without explicitly coordinating with execution-layer intelligence. Similarly, rApps operating in the Non-RT RIC provide powerful capabilities for long-term policy design and AI-driven network optimization. However, rApps alone cannot react to fast-changing radio conditions or enforce resource allocation decisions at fine time granularity [ 10 ] . In addition, the architectural introduction of dApps in the O-RAN control loops enables direct radio control in sub-slot or even TTI-level timescales, aligning with the latency requirements of scheduling and far-edge radio resource allocation [ 4 , 14 ] . Hence, the adoption of dApps opens new opportunities for hierarchical control loops in which long-term policies (defined by rApps) and interference-aware resource allocation (performed by xApps) can be complemented by ultra-fast scheduling and resource arbitration mechanisms at the DU level. Despite these emerging capabilities, existing O-RAN optimization frameworks largely treat rApps, xApps, and execution-layer schedulers at the DU as independent control entities. Coordinating these layers introduces several challenges, including how to distribute optimization tasks across different time scales for end-to-end optimization, how to maintain stability across nested control loops, and how to exchange information between policy, control, and execution layers without introducing excessive signaling overhead. Furthermore, relying solely on rApp- or xApp-based control not only concentrates optimization decisions at the Non- or Near-RT RIC, often leading to frequent O1 or E2 signaling and potential interface overload, but also limits decision responsiveness to minute- or second-level timescales [ 7 ] . In contrast, a coordinated and nested rApp–xApp–dApp architecture can decouple control responsibilities across layers, balance communication overhead, and enable responsive resource management across multiple time scales [ 4 ] . To address this gap, this article proposes CollabORAN, a collaborative spatiotemporal spectrum management framework that enables coordinated rApp–xApp–dApp control loops for equitable spectrum access in O-RAN networks. By decoupling the fair spectrum allocation problem into different timescale-specific tasks, the proposed architecture leverages a hierarchical O-RAN structure to separate an AI/ML-aided and SLA-driven policy definition (through AI/ML traffic forecasting), interference-aware spectrum allocation (through dynamic hypergraph construction and coloring), and resource time-sharing scheduling across different O-RAN layers and time scales. The specific contributions of CollabORAN are summarized as follows: • The framework provides generalized collaborative multi-timescale rApp–xApp–dApp control that enables end-to-end policy-driven dynamic spectrum management in O-RAN networks. Policies at the Non-RT RIC are dynamically co-configured by the operator-defined SLAs and AI/ML-assisted traffic forecasts from the rApp (called FrerApp). • A graph-based interference-aware and mobility-aware physical resource block (PRB) allocation mechanism (called FrexApp) is implemented at the xApp layer to dynamically construct interference conflict graphs and allocate spectrum resources across users covered by heterogeneous RUs. • A graph-constrained modified proportional-fair (PF) temporal scheduling mechanism (called FredApp) is applied at the dApp layer, enabling dynamic PRB time sharing that prevents long-term user starvation and improves fairness. • A comprehensive and generalizable workflow analysis of the proposed hierarchical and collaborative closed loop is presented. In addition, performance evaluation in heterogeneous O-RAN environments is conducted to demonstrate the CollabORAN performance in terms of traffic forecasting, interference mitigation and service share balance, under dynamic traffic conditions. The results are compared against standalone deployments of intelligent applications to showcase the tangible potency of CollabORAN. II CollabORAN Architectural Overview II-A Design Principles The design of CollabORAN is guided by three key architectural principles that enable scalable and coordinated spectrum management across the O-RAN ecosystem. Principle 1 - Separation of concerns

Read the complete text