Enabling SLO-Aware 5G Multi-Access Edge Computing with SMEC
summary
SMEC implements decoupled RAN and edge schedulers that infer deadlines from existing signals—5G control signals at the RAN and a lightweight probe-and-ack probing plus application life‑cycle events at the edge—so no RAN–edge coordination or app rewrites are needed. Prototype runs on a 5G testbed (SR 5G stack), Intel CPU and Nvidia L4 GPU, with three UDP apps. Evaluation reports 90–96% SLO satisfaction versus under 6% for prior approaches and up to 122× tail‑latency reduction.
AI-written summary · about
chapters (14)
- 0:00SLO-aware MEC introduction
- 0:40MEC and SLOs background
- 2:03High latency variability in MEC
- 3:25Latency variability in 5G
- 5:34Limitations of SLO-unaware schedulers
- 7:00AsMAC: SLO-aware framework
- 8:25Inferring timing for SLOs
- 10:10Recovering missing 5G signals
- 11:57Network latency via downlink stability
- 13:43Lightweight delay probing
- 14:05Request lifecycle timing
- 15:10AsMAC system integration
- 15:51AsMAC evaluation
- 16:57Summary and future MEC
related talks
12mP4EC: Enabling Terabit Edge Computing in Enterprise 4G LTE
01h 25mDay 2 Part 1 Sessions
11mA Comparison of Communication Mechanisms in Vehicular Edge Computing
44mHow 5G Edge is being used to solve customer problems
25mKeynote: Network Transformation: Are We Ready for 5G and Edge Clouds?
24mCisco 5G Automated Deployment
17mEdgeRIC: Empowering Real-time Intelligent Optimization and Control in NextG Cellular...
22mDefensive 5G