Indoor Sharing in the Mid-Band: A Testbed Evaluation of Neutral-Host, Cellular Macro & Wi-Fi

Author(s)

Joshua Roy Palathinkal, Muhammad Iqbal Rochman, Vanlin Sathya, Mehmet Yavuz, and Monisha Ghosh

Abstract

Indoor environments present significant challenges for wireless connectivity. Public Mobile Network Operators (MNOs) utilizing outdoor macro base stations (BSs) to serve indoor customers suffer from poor signal penetration while indoor Wi-Fi networks may face reliability issues due to spectrum contention. Shared spectrum models, particularly the Citizens Broadband Radio Service (CBRS) band utilized by private 4G/5G networks, have emerged as a promising alternative to providing reliable indoor service. Moreover, these private networks are equipped with the neutral-host (NH) model, seamlessly offloading indoor MNOs’ traffic to the private CBRS network. This paper presents a comprehensive, real-world performance evaluation of three co-located technologies utilizing mid-band spectrum (1-6 GHz): a CBRS-based NH network, public MNO macro networks, and a Wi-Fi 6 network, inside a large, big-box retail store characterized by significant building loss. While the evaluation is site-specific, it is a real-world representation of a highly dense class of indoor deployments. Our analysis demonstrates: (i) the NH network provides superior indoor coverage compared to MNO macro, requiring only six CBRS devices (CBSDs) versus 65 Access Points (APs) for enterprise Wi-Fi to achieve full coverage, with a median building loss of 26.6 dB ensuring interference-free coexistence with outdoor federal incumbents; (ii) the NH network achieves substantial indoor throughput gains, with per-channel normalized throughput improvements of 1.44x and 1.62x in downlink (DL), and 4.33x and 13x in uplink (UL), compared to 4G and 5G macro deployments, respectively; (iii) the NH deployment achieves a median indoor aggregated physical (PHY)-layer DL throughput gain of 2.08x over 5G macro deployments indoors, despite utilizing only 40 MHz of aggregated bandwidth compared to 225 MHz for 5G macro; and (iv) the NH deployment also outperforms Wi-Fi in application-layer HTTP DL performance by 5.05x. The findings offer critical insights by presenting: (i) a framework to leverage real-world deployments as testbeds for studying indoor shared-spectrum networks; and (ii) measurement-based evidence supporting the use of NH to improve indoor cellular coverage.