One of the presumed outcomes of the 5G process is full convergence of licensed and unlicensed spectrum, with one or more air interface standards which can span both, using frequencies entirely flexibly according to requirement. This is a very long way off, if the current quarrels over extending LTE into licence-exempt bands are anything to go by.
Coexistence of Wi-Fi and LTE in the 5 GHz band is beset by political and commercial agendas, but there are also genuine technical problems in the way of being good neighbors. This is unsurprising, given that the two technologies started from very different places and have evolved separately, under the auspices of different standards bodies. Is converged, evolutionary 5G really practical?
Will this situation improve in the 5G era? The 3GPP is in control of the air interface standards process again and there seems to be little input from Wi-Fi’s IEEE body, though most Wi-Fi players expect their technology to evolve into a 5G component too. Operators want 5G to be an evolution of LTE, not a rip-and-replace new start, but the dream of a fully converged licensed/unlicensed solution might be more achievable with a new platform, avoiding the issues LTE-Unlicensed is now facing – many arising from the difficulties of forcing two such different wireless standards to play nicely together.
And that’s before the regulatory and ownership issues of convergence are aired, and the impact on the value of spectrum. Again, a clean slate would be desirable, to usher in a new approach to spectrum holdings, based around flexible access and doing away with outdated auctions and multi-decade, multi-million-dollar licences. But treasuries and spectrum owners alike will protest and the result, like LTE-U itself, is likely to be an uneasy compromise between the entrenched positions and the radical vision.
It is sadly easy to see, in the current 5 GHz rows, glimpses of the likely debates which will risk delaying or derailing full 5G platforms in the 2020s. The latest outbreak of hostility over LTE-U enters on a testing plan, proposed by the Wi-Fi Alliance, and more specifically over signal strength thresholds. LTE-U is, of course, particularly controversial because it does not have to support listen before talk (LBT), which is required for coexistence in the 5 GHz band in many parts of the world, but not in the US. LTE-U uses carrier sensing adaptive transmission (CSAT) instead.
Other unlicensed LTE solutions, such as LTE-LAA (licensed assisted access) and Qualcomm’s MuLTEfire, do implement LBT (so could be used worldwide if regulators agree) and have other safeguards for Wi-Fi. But those are yet to be commercialized (LTE-LAA is part of the current 3GPP Release 13 but will not be readily available until 2017), while LTE-U was part of Releases 10 to 12 and so could be rolled out now in the US and some other countries, if it can get the regulatory green light.
The Wi-Fi Alliance’s test regime draws backlash:
Hence the urgency with which the Wi-Fi community has claimed interference risks to try to block LTE-U, while also insisting on taking the lead in devising a coexistence framework, should the blocking effort fail. Meanwhile, many cellular players are concerned about the behaviour of Wi-Fi near their LTE signals - LTE-U adapts its duty cycle to coexist with Wi-Fi, and treats Wi-Fi devices as slaves, while Wi-Fi devices occupy the whole airtime, for instance.
The Wi-Fi Alliance has published its proposed test plan, but received a furious response from Qualcomm and others (even though Qualcomm is a prominent Wi-Fi player since its acquisition of Atheros, it is a cellular organization at heart, and its leadership of several 5 GHz LTE initiatives, from LTE-U commercialization to MuLTEfire, show how the chip giant continues to assert engineering and IPR dominance in the cellular platform, in a way it has never achieved in Wi-Fi.
The main issue around acceptable testing parameters concerns Wi-Fi received signal strength (RSSI) values. The Alliance has specified -82 dBm, but Verizon and Qualcomm, among others, believe the value should be -72 dBm. The Qualcomm camp believes the -82 dBm test is too rigorous, since Wi-Fi, Bluetooth and ZigBee only back off at levels of -62 dBm. LAA has already settled on a compromise level of -72 dBm for testing (but, of course, supports LBT, making the issue less divisive). But Broadcom and other Wi-Fi players say the tests should ideally be more demanding, testing signal sensitivity scenarios as low as -89 dBm.
“The latest version of the test plan released by the Wi-Fi Alliance lacks technical merit, is fundamentally biased against LTE-U, and rejects virtually all the input that Qualcomm provided for the last year,” Dean Brenner, Qualcomm’s SVP of government affairs, said in a statement earlier this month. “The latest version of the plan would require LTE-U to protect Wi-Fi 100 times more than Wi-Fi would protect LTE-U in all environments under criteria that ignore data submitted to the Wi-Fi Alliance, including data from Wi-Fi vendors.”
He expressed anger that the Alliance had rejected a compromise proposal submitted by Qualcomm and claimed the new plan had “the clear purpose of trying to keep the benefits of LTE-U away from consumers and off the unlicensed spectrum, which is supposed to be for all of us”. He added: “The watchword for unlicensed spectrum is supposed to be permissionless innovation, not incumbent protection,” a conclusion which had a certain irony, coming from the camp whose key interest is to defend the position of incumbents like Verizon.
Brenner urged the FCC to “disregard” the Alliance’s proposal, hinting that the regulator may have to intervene to move things forward – something it has been reluctant to do in unlicensed spectrum.
Copyright © 2016, Wireless Watch
Wireless Watch is published by Rethink Research, a London-based IT publishing and consulting firm. This weekly newsletter delivers in-depth analysis and market research of mobile and wireless for business. Subscription details are here.