In-Network Artificial Computing Enhanced Light Model-Switching for Emergency Communications Networks

TL;DR

This paper presents a lightweight in-network artificial computing framework enabling rapid model switching for emergency networks, achieving high throughput and low latency on commodity hardware.

Contribution

It introduces a novel in-network model-switching approach with efficient packet-level selection and demonstrates its practicality on standard hardware.

Findings

Achieves 1.894 million packets per second throughput.

Maintains 0.528 microseconds inference latency.

Supports 16 resident models with minimal switching overhead.

Abstract

Emergency communications networks require in-network intelligence for timely traffic handling under dynamic demands and runtime constraints. In these environments, packets may need different inference behaviors, and conventional model replacement via control-plane updates is too slow for responsive operation. We propose an in-network artificial computing framework with lightweight model-switching, where multiple Binary Neural Network (BNN) models are kept resident within a shared execution framework. Packet metadata selects the active model at packet granularity with O(1) selection cost. A fixed 1024-byte payload is aligned with x86 AVX-512, enabling efficient memory access. The framework is realized on an eBPF/XDP + AF_XDP stack. Experimental results show that the system sustains 1.894 Mpps with a 0.528 us inference latency, while model selection adds only 0.005 us. Our results demonstrate that different resident models induce distinct packet-processing behaviors, that scaling to 16 slots preserves low switching overhead, and that online model switching completes without wrong-verdict packets. These results show the practicality of lightweight in-network artificial computing on commodity hardware.