O-RAN End-to-End performance metrics

This dataset is related to the GNN-based E2E Performance Prediction in 6G-SENSES deliverable D5.2 (section 5.1). It considers x-haul network traffic, generated by end-nodes: RAN, sensing and core elements, as well as configurable traffic forwarding nodes. The generated traffic flows belong to both communication and sensing services. In the case of communication flows, these are: fronthaul traffic according to the underlying RAN configuration; backhaul traffic following specific service features (i.e. URLLC, eMMB, mMTC); sensing traffic according to specific sensing capability configuration; in the following it will be referred as sensing backhaul. The forwarding nodes allow different labelling and traffic scheduling policies. The dataset records end-to-end (E2E) metrics of the different flows that can be used to compute per-packet delay, jitter, and throughput. In addition, sojourn time at intermediate nodes is also recorded.本数据集与6G-SENSES项目交付件D5.2（第5.1节）中基于图神经网络（Graph Neural Network，GNN）的端到端（End-to-End，E2E）性能预测研究相关。其涵盖由终端节点生成的x-haul网络流量，终端节点包括无线接入网（Radio Access Network，RAN）、感知节点与核心网元，以及可配置流量转发节点。所生成的流量流同时涵盖通信与感知两类业务。就通信流量流而言，其包含：基于底层无线接入网配置的前传流量；遵循特定业务特性的回传流量（即超可靠低延迟通信（Ultra-Reliable Low Latency Communications，URLLC）、增强移动宽带（enhanced Mobile Broadband，eMMB）、海量机器类通信（massive Machine Type Communications，mMTC））；以及基于特定感知能力配置的感知流量；下文将此类感知流量称为感知回传流量。转发节点支持多种流量标记与调度策略。本数据集记录了各类流量流的端到端指标，可用于计算单包延迟、抖动与吞吐量。此外，还记录了流量在中间节点的驻留时间。