Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

PACIFISTA: Conflict Evaluation and Management in Open RAN

Pietro Brach del Prever, Salvatore D'Oro, Leonardo Bonati, Michele Polese, Maria Tsampazi, Heiko Lehmann, Tommaso Melodia

TL;DR

Open RAN enables multiple autonomous AI-based controllers (rApps/xApps/dApps) whose competing goals can cause instability and performance degradation. PACIFISTA provides a data-driven, graph-based framework that (i) profiles O-RAN applications in sandbox environments, (ii) detects direct, parameter, and KPI conflicts using augmented parameter and KPI graphs, (iii) evaluates conflict severity with distribution-based distance metrics, and (iv) mitigates conflicts via a threshold-based deployment policy. Core contributions include formal definitions of conflict classes, G^P and G^K graphs, a profiling pipeline that yields per-condition statistical profiles (with at least 3,000 samples per case), a comprehensive conflict report with severity indices, and a practical mitigation strategy demonstrated on Colosseum/OpenRAN Gym. The approach enables pre-deployment risk assessment and dynamic, policy-driven co-existence of diverse xApps/rApps, enhancing stability and operator control in open, multi-vendor RAN environments.

Abstract

The O-RAN ALLIANCE is defining architectures, interfaces, operations, and security requirements for cellular networks based on Open Radio Access Network (RAN) principles. In this context, O-RAN introduced the RAN Intelligent Controllers (RICs) to enable dynamic control of cellular networks via data-driven applications referred to as rApps and xApps. RICs enable for the first time truly intelligent and self-organizing cellular networks. However, enabling the execution of many Artificial Intelligence (AI) algorithms making autonomous control decisions to fulfill diverse (and possibly conflicting) goals poses unprecedented challenges. For instance, the execution of one xApp aiming at maximizing throughput and one aiming at minimizing energy consumption would inevitably result in diametrically opposed resource allocation strategies. Therefore, conflict management becomes a crucial component of any functional intelligent O-RAN system. This article studies the problem of conflict mitigation in O-RAN and proposes PACIFISTA, a framework to detect, characterize, and mitigate conflicts generated by O-RAN applications that control RAN parameters. PACIFISTA leverages a profiling pipeline to tests O-RAN applications in a sandbox environment, and combines hierarchical graphs with statistical models to detect the existence of conflicts and evaluate their severity. Experiments on Colosseum and OpenRAN Gym demonstrate PACIFISTA's ability to predict conflicts and provide valuable information before conflicting xApps are deployed on production. We demonstrate that users can experience a 16% throughput loss even in the case of xApps with similar goals, and that applications with conflicting goals might cause instability and result in up to 30% performance degradation. We also show that PACIFISTA can help operators to identify conflicting applications and maintain performance degradation at bay.

PACIFISTA: Conflict Evaluation and Management in Open RAN

TL;DR

Open RAN enables multiple autonomous AI-based controllers (rApps/xApps/dApps) whose competing goals can cause instability and performance degradation. PACIFISTA provides a data-driven, graph-based framework that (i) profiles O-RAN applications in sandbox environments, (ii) detects direct, parameter, and KPI conflicts using augmented parameter and KPI graphs, (iii) evaluates conflict severity with distribution-based distance metrics, and (iv) mitigates conflicts via a threshold-based deployment policy. Core contributions include formal definitions of conflict classes, G^P and G^K graphs, a profiling pipeline that yields per-condition statistical profiles (with at least 3,000 samples per case), a comprehensive conflict report with severity indices, and a practical mitigation strategy demonstrated on Colosseum/OpenRAN Gym. The approach enables pre-deployment risk assessment and dynamic, policy-driven co-existence of diverse xApps/rApps, enhancing stability and operator control in open, multi-vendor RAN environments.

Abstract

The O-RAN ALLIANCE is defining architectures, interfaces, operations, and security requirements for cellular networks based on Open Radio Access Network (RAN) principles. In this context, O-RAN introduced the RAN Intelligent Controllers (RICs) to enable dynamic control of cellular networks via data-driven applications referred to as rApps and xApps. RICs enable for the first time truly intelligent and self-organizing cellular networks. However, enabling the execution of many Artificial Intelligence (AI) algorithms making autonomous control decisions to fulfill diverse (and possibly conflicting) goals poses unprecedented challenges. For instance, the execution of one xApp aiming at maximizing throughput and one aiming at minimizing energy consumption would inevitably result in diametrically opposed resource allocation strategies. Therefore, conflict management becomes a crucial component of any functional intelligent O-RAN system. This article studies the problem of conflict mitigation in O-RAN and proposes PACIFISTA, a framework to detect, characterize, and mitigate conflicts generated by O-RAN applications that control RAN parameters. PACIFISTA leverages a profiling pipeline to tests O-RAN applications in a sandbox environment, and combines hierarchical graphs with statistical models to detect the existence of conflicts and evaluate their severity. Experiments on Colosseum and OpenRAN Gym demonstrate PACIFISTA's ability to predict conflicts and provide valuable information before conflicting xApps are deployed on production. We demonstrate that users can experience a 16% throughput loss even in the case of xApps with similar goals, and that applications with conflicting goals might cause instability and result in up to 30% performance degradation. We also show that PACIFISTA can help operators to identify conflicting applications and maintain performance degradation at bay.
Paper Structure (27 sections, 4 equations, 18 figures, 7 tables)

This paper contains 27 sections, 4 equations, 18 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Modeling Conflicts in O-RAN
  4. Definitions of Conflicts
  5. PACIFISTA's Conflict Model
  6. PACIFISTA Overview
  7. PACIFISTA in a Nutshell
  8. Integration with O-RAN
  9. Profiling O-RAN Applications
  10. Creating Statistical Profiles
  11. Detecting conflicts
  12. Detecting Direct Conflicts
  13. Detecting Parameter Conflicts
  14. Detecting Conflicts
  15. Evaluating conflicts
  16. ...and 12 more sections

Figures (18)

  • Figure 1: Impact of conflicts on performance. The xApps for and for are first run separately and then together. Left: assigned for slice. Right: Measured throughput statistics.
  • Figure 2: PACIFISTA architecture and workflow.
  • Figure 3: Examples of conflicts and graphs used in PACIFISTA.
  • Figure 4: Application profile comparison for different number of samples for slice of xApp $a_5$.
  • Figure 5: Profiling of new O-RAN applications.
  • ...and 13 more figures

Theorems & Definitions (5)

  • Definition 1: Direct Conflict
  • Definition 2: Parameter Conflict
  • Definition 3: Conflict
  • Remark 1
  • Remark 2