NTP-INT: Network Traffic Prediction-Driven In-band Network Telemetry for High-load Switches

TL;DR

This work addresses the challenge of acquiring fine_grained in_band telemetry from high_load switches in dynamic 6G-like networks without incurring prohibitive overhead. It introduces NTP-INT, a three_module system that combines a Multi_Temporal Graph Neural Network MTGNN for traffic_prediction with a network_pruning stage to generate a compact subnetwork, and an attention_based DRL probe_path_planner to optimize high_frequency telemetry under latency constraints. Key contributions include integrating MTGNN with INT, designing subconnected_graph generation with articulation_point_detection and biconnected_graph generation for robust slices, and developing a masking_enabled DRL framework that significantly reduces training_time and control_overhead (≈50%). The results demonstrate improved telemet ry accuracy for high_load_switches and substantially lower overhead, enabling scalable, real_time visibility in dynamic high_load_networks.

Abstract

In-band network telemetry (INT) is essential to network management due to its real-time visibility. However, because of the rapid increase in network devices and services, it has become crucial to have targeted access to detailed network information in a dynamic network environment. This paper proposes an intelligent network telemetry system called NTP-INT to obtain more fine-grained network information on high-load switches. Specifically, NTP-INT consists of three modules: network traffic prediction module, network pruning module, and probe path planning module. Firstly, the network traffic prediction module adopts a Multi-Temporal Graph Neural Network (MTGNN) to predict future network traffic and identify high-load switches. Then, we design the network pruning algorithm to generate a subnetwork covering all high-load switches to reduce the complexity of probe path planning. Finally, the probe path planning module uses an attention-mechanism-based deep reinforcement learning (DEL) model to plan efficient probe paths in the network slice. The experimental results demonstrate that NTP-INT can acquire more precise network information on high-load switches while decreasing the control overhead by 50\%.

Figures (15)

• Figure 1: Architecture of NTP-INT.
• Figure 2: Workflow of NTP-INT.
• Figure 3: The framework of the MTGNN.
• Figure 4: Graph convolution module
• Figure 5: Temporal convolution module