Towards Practical and Usable In-network Classification

2025-12-10

AI Summary

Existing in-network machine learning (ML) approaches face practical deployment barriers due to scarce hardware resources, complex implementation, and low developer accessibility. To address these challenges, this paper introduces ACORN—a system enabling runtime-programmable deployment of decision trees, random forests, and support vector machines (SVMs) on P4-programmable switches. ACORN proposes a lightweight data-plane ML representation, an end-to-end automated translation and optimization toolchain from Python ML models to P4, and an integer linear programming (ILP)-based co-design framework for joint feature selection and resource allocation. Experimental evaluation demonstrates that ACORN achieves near-lossless performance—throughput and latency degradation remain under 3%—while supporting 2×–4× more features than prior work, significantly enhancing model capacity and deployment efficiency. The complete source code and toolchain will be open-sourced.

Technology Category

• Machine Learning: Hardware-aware ML
• Humans and AI: Human-in-the-loop Machine Learning
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

In-network machine learning enables real-time classification directly on network hardware, offering consistently low inference latency. However, current solutions are limited by strict hardware constraints, scarce on-device resources, and poor usability, making them impractical for ML developers and cloud operators. To this end, we propose ACORN, an end-to-end system that automates the distributed deployment of practical machine learning models across the network. ACORN provides a fully automated pipeline that loads and deploys Python ML models on network devices using an optimized deployment plan from an ILP planner. To support larger models under hardware constraints and allow runtime programmability, ACORN adopts a novel data plane representation for Decision Tree, Random Forest, and Support Vector Machine models. We implement ACORN prototype in P4 and run it on real programmable hardware. Our evaluation shows ACORN can deploy classification ML models with 2-4x more features than state-of-the-art solutions, while imposing negligible overhead on network performance and traffic. We will make our data plane program, model translator, optimizer, and all related scripts publicly available.

Problem

• Automates deployment of ML models on network hardware
• Overcomes hardware constraints for in-network classification
• Enables larger models with minimal network performance impact

Innovation

• Automated pipeline for Python ML models on network devices
• Novel data plane representation for Decision Tree, Random Forest, SVM
• Deploys models with 2-4x more features than current solutions