MAC-Gyver: Open, Programmable, Scheduling for AI-RAN 6G Systems

2026-07-28

AI Summary

This work addresses the challenge of deploying and validating learning-based wireless schedulers in real systems, hindered by the absence of an open, programmable, and end-to-end evaluable platform. To bridge this gap, we present MAC-Gyver—the first open-source framework tailored for production-grade 5G/6G scheduling stacks. MAC-Gyver extends the OpenAirInterface Layer 2 scheduler with typed interfaces, introduces a PHY-less real-time emulator (mac-emu), and integrates 3GPP-compliant channel models with PRB-level frequency-selective scheduling, thereby unifying simulation, control, and over-the-air testing environments. The framework enables efficient iteration and 3GPP-compatible validation of AI-driven scheduling policies. Experimental results demonstrate that proactive uplink scheduling reduces median round-trip latency by nearly 50%, while frequency-selective uplink scheduling consistently approaches offline performance bounds across diverse mobility and power constraints.

Abstract

Cellular networks are integrating Artificial Intelli- gence (AI) into radio access network control. The MAC scheduler is a promising target because it allocates a limited resource, spectrum, at every slot, under competing latency, throughput, and reliability requirements. However, most learning-based sched- ulers are evaluated only in simulation. Production schedulers are difficult to modify, and realistic stress tests require more radio hardware than most laboratories can provide. We present MAC-Gyver, an open-source framework for developing and evaluating scheduling applications that execute directly inside the OpenAirInterface scheduler. It exposes scheduler observations and controls through typed interfaces while preserving the underlying protocol and real-time execution paths. The same applications run over the air and in mac-emu, a PHY-less emulator that executes the unmodified OpenAirInterface Layer 2 stack for up to 90 users on one host at real-time slot pace, with a 3GPP-compliant channel model. To showcase the flexibility of MAC-Gyver, we evaluate two use cases. A proactive uplink scheduler predicts packet arrivals and roughly halves median round-trip latency. A frequency-selective uplink scheduler selects contiguous sub-bands from per-PRB sounding observations and is evaluated across mobility and power-limited operating points against an offline scheduling ceiling. Together, they show how the same production stack can be an AI playground that supports implementation, controlled evaluation, and over-the-air validation through complementary scheduling use cases.

Problem

• AI-RAN
• MAC scheduler
• 6G systems
• real-time evaluation
• open-source framework

Innovation

• AI-RAN
• MAC scheduling
• OpenAirInterface
• real-time emulation
• programmable scheduler