Channel-Aware 5G RAN Slicing with Customizable Schedulers
summary
RadioSaber implements channel-aware RAN slicing by making both inter-slice and enterprise schedulers greedy inter-slice and enterprise schedulers so the inter-slice allocator queries each slice: “who would take this RB?” and assigns RBs to maximize per-RB channel rates. It runs on an open-source 5G RAN simulator with LTE-scoped channel traces, evaluates up to 800 users and 20 slices (5–15 users each), enforces weighted isolation, has low runtime overhead, and reports 17–72% throughput gain over channel-agnostic inter-slice approaches.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 0:32Network slicing basics
- 2:59Need for channel awareness
- 6:56The cyclic dependency challenge
- 9:05Radio-Saber architecture
- 11:36Enterprise scheduler design
- 13:32Performance evaluation
- 15:41Conclusion
related talks
17mApplication-Level Service Assurance with 5G RAN Slicing
18mEnabling SLO-Aware 5G Multi-Access Edge Computing with SMEC
17mTowards Energy Efficient 5G vRAN Servers
18mCHISEL: An optical slice of the wide-area network
21mE2E 5G network slicing in ONAP
17mDesigning Next Generation 5G Wireless Infrastructure Using RISC-V
51mAn Account of OpenAirInterface’s Open Software in 5G RAN, CORE NETWORK and MOSAIC5G
21mKeynote: Re-defining the Competitive Landscape