How Much Cache is Needed to Achieve Linear Capacity Scaling in Backhaul-Limited Dense Wireless Networks?
An Liu, Vincent Lau

TL;DR
This paper investigates the minimum cache size required at base stations in dense wireless networks to achieve linear capacity scaling, demonstrating that sufficient caching can replace wired backhauls and significantly improve network throughput.
Contribution
It introduces a novel backhaul-limited cached dense wireless network architecture and quantifies the cache size needed for linear capacity scaling and MIMO cooperation gains.
Findings
Capacity scales linearly with the number of base stations if cache size exceeds a content-dependent threshold.
Cache-induced MIMO cooperation provides significant throughput gains over traditional caching schemes.
Minimum cache size for MIMO gain is the same as that for linear capacity scaling.
Abstract
Dense wireless networks are a promising solution to meet the huge capacity demand in 5G wireless systems. However, there are two implementation issues, namely the interference and backhaul issues. To resolve these issues, we propose a novel network architecture called the backhaul-limited cached dense wireless network (C-DWN), where a physical layer (PHY) caching scheme is employed at the base stations (BSs) but only a fraction of the BSs have wired payload backhauls. The PHY caching can replace the role of wired backhauls to achieve both the cache-induced MIMO cooperation gain and cache-assisted Multihopping gain. Two fundamental questions are addressed. Can we exploit the PHY caching to achieve linear capacity scaling with limited payload backhauls? If so, how much cache is needed? We show that the capacity of the backhaul-limited C-DWN indeed scales linearly with the number of BSs if…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
