POSMAC: Powering Up In-Network AR/CG Traffic Classification with Online Learning

2025-02-02

AI Summary

To address the trade-off between latency and throughput in real-time traffic classification for augmented reality (AR) and cloud gaming (CG) applications—where conventional CPU/GPU-based approaches fall short—this paper proposes a hardware-accelerated online learning framework built on NVIDIA DOCA-enabled Data Processing Units (DPUs). The method natively deploys lightweight decision trees and random forests on ARM-based DPUs, enabling dynamic model updates and low-overhead generalization evaluation, thereby achieving the first DPU-native online learning implementation for network traffic classification. Compared to software-only solutions, the system attains near line-rate throughput (>100 Gbps), end-to-end latency under 100 μs, and significantly improved AR/CG classification accuracy. The core contribution lies in the tight co-design of online learning algorithms and DPU hardware acceleration, effectively overcoming performance bottlenecks in real-time, in-network traffic classification.

Technology Category

• Machine Learning: Hardware-aware ML
• Search and Optimization: Learning to Search
• Computer Vision: Learning & Optimization for CV

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

In this demonstration, we showcase POSMAC1, a platform designed to deploy Decision Tree (DT) and Random Forest (RF) models on the NVIDIA DOCA DPU, equipped with an ARM processor, for real-time network traffic classification. Developed specifically for Augmented Reality (AR) and Cloud Gaming (CG) traffic classification, POSMAC streamlines model evaluation, and generalization while optimizing throughput to closely match line rates.

Problem

• AR and CG Traffic Classification
• Real-time Judgment Capability
• Network Performance Optimization

Innovation

• Real-time Decision Tree
• Random Forest Model
• ARM Processor on DPU