Much Ado: Assessing Wi-Fi/LTE Coexistence in Unlicensed Bands
45 Pages Posted: 27 Mar 2016 Last revised: 27 Oct 2016
Date Written: March 30, 2016
Spectrum allocation tussles between incumbents and new entrants have traditionally occurred between licensed services in adjacent bands. By contrast, the recent fight over the deployment of LTE in the unlicensed bands (LTE-in-Unlicensed) addresses services that operate on the same frequencies and are not entitled to interference protection from each other.
We present a study of likely LTE-in-Unlicensed entrant systems coexisting with incumbent IEEE 802.11n Wi-Fi in the 5 GHz unlicensed band. We explore a large design parameter space for several distinct entrant technologies (e.g. LTE-U, LAA) in several deployment scenarios, for different incumbent and entrant node densities. We evaluate the coexistence performance in terms of the downlink throughput per entrant/incumbent access point, based on Monte Carlo simulations in MATLAB.
Our work was not commissioned to serve an advocacy purpose, and our results contradict some of the claims of both camps. Key technical findings bearing on the public policy debate are: (1) For realistic node densities, near-perfect coexistence of Wi-Fi and LTE-in-Unlicensed is ensured by the large number of 5 GHz channels, so that co-channel operation is easily avoided. (2) Evaluated by Wi-Fi throughput, LTE-U and LAA entrants are almost equally good neighbours to Wi-Fi incumbents. (3) For the worst-case scenario of forced co-channel operation – or equivalently, extremely high node densities – LTE-in-Unlicensed is occasionally a better neighbour to Wi-Fi than a Wi-Fi entrant, but sometimes worse.
We examine previous cases of tussles between unlicensed services, as a policy context for our LTE-in-Unlicensed/Wi-Fi coexistence study: HomeRF/Wi-Fi, Blue-tooth/Wi-Fi, and IEEE 802.11b vs. 802.11g. We find that cooperation among engineers from competing parties can significantly mitigate coexistence problems. Also, some debates were rendered moot when one of the competing technologies became dominant in the market, or they were deployed together in devices.
We conclude that if a regulator chooses to use entrant/incumbent coexistence as a criterion for rulemaking in unlicensed bands (which we posit here, but do not endorse), then LTE/Wi-Fi coexistence in 5 GHz does not require any action since problems will be rare. If the regulator chooses to over-weight problematic edge cases, then again there is no basis for action since LTE-in-Unlicensed is some-times friend, and sometimes foe, to Wi-Fi. We submit that the regulator should demand a systematic engineering analysis, surveying a wide range of parameters and scenarios, for supporting evidence-based rulemaking.
Keywords: allocation, coexistence, interference, LTE, spectrum, unlicensed, Wi-Fi
Suggested Citation: Suggested Citation