The promise of the USA’s multi-tiered CBRS spectrum band has been discussed for so long that it is easy to forget that there are still no commercial services in the 3.5 GHz band, either in its licensed or shared tiers; and still barriers to full deployment, though most of them are political rather than technical. And with the timing of 5G in shared spectrum uncertain, there is likely to be plenty of time for CBRS to make its mark on the mobile services landscape, even if impatience is rising for it to kick off its activities.
So much is riding on the success of the FCC’s bold experiment with opening up a formerly federal band for mobile broadband, with three different levels of priority access – guaranteed top priority for federal incumbents; followed by holders of licences (Priority Access Licensing or PAL); and then by General Authorized Access or GAA – unlicensed, subject to being assigned vacant channels by a spectrum allocation system (SAS), which will manage all the priority levels. Round the world, supporters of a more open ecosystem of mobile service providers, including industrial specialists, are hoping that CBRS will be sufficiently successful to encourage similar schemes elsewhere.
Most focus on the potential opportunity for new mobile entrants has focused on the general access tier, but PAL could also encourage new players, because the licences are likely to be shorter term, and more localized, than most mainstream mobile franchises, and so more affordable for non-traditional MNOs. Wireline or industrial providers have shown interest, since they might want more assured performance and capacity than the shared system would support, but of course, the mobile operators are also keen to snap up the airwaves, to add affordable capacity to their networks, and to fend off those new players.
As commercial reality draws closer, there has been plenty of activity in tests, trials and new equipment, as well as intensifying political arguments – many over how short and small the PAL licences should be, with the MNOs lobbying for longer terms, larger territories and higher barriers to newcomers; while web giants, industrial companies like GE and cablecos have led the charge for a more flexible system.
Last week, the newest in a series of tests was revealed via an application by Google – a prolific triallist and ecosystem promoter in CBRS, as well as the holder of one of the SAS licences. This latest application, for a Special Temporary Authority (STA) to conduct radio tests in the 3550-3650 MHz band, does not give much away. The tests would take place over six months in Philadelphia, starting on October 25, and would trial transmission of broadband data to mobile devices from fixed locations, but beyond that generic description, there is little more that hasn’t been redacted in the document.
Google says it needs to keep the information confidential because it has “significant commercial value”, and because of the non-disclosure agreements it holds with equipment suppliers. “The manufacturer name and model number constitute confidential trade secrets, technical information and business information under the agreements,” Google says in its filing.
Meanwhile, Google has received conditional approval as an SAS administrator and says it will harness this capability to facilitate an end-to-end CBRS platform, working with vendors, service providers, incumbents and other SAS administrators. Those other conditionally approved SAS providers are Federated Wireless – which has already conducted SAS interoperability tests with Google – plus CommScope, Sony, Amdocs and Key Bridge Wireless.
An end-to-end system, under the control of an expert orchestrator, could achieve the kind of platform Google has been discussing for a decade – a wholesale, neutral host framework hosting large numbers of MVNOs or enterprise service providers, on a flexible, on-demand basis, without the rigidity and high cost of traditional MVNO models. In the USA, many vertical industries have expressed interest in such a system, which would enable them to harness shared spectrum without the cost of buying licences (though others, as noted above, are also interested in PALs to support higher predictability of service).
Because there is 150 MHz of spectrum potentially available for sharing in CBRS, there is, for the first time outside the WiFi world, sufficient capacity to support large numbers of players and use cases, without long term spectrum or MVNO commitment. Educational establishments, transport providers, healthcare providers, stadiums and retailers are among the sectors which expressed interest in using CBRS for private LTE, during the FCC consultations.
The FCC recently asked for details from the would-be SAS administrators about their initial commercial deployment plans. Google said its SAS could support service anywhere in the USA, but in reality, it expects initial deployments to be discrete and focused on particular locations, relatively limited in size. By contrast, Federated says it will turn on CBRS spectrum services across 15,773 sites in 47 states and the District of Columbia.
As a relatively high frequency band with power limitations, the shared access portion of CBRS will be best suited to small cells and to ‘hotzones’ of urban or rural coverage and capacity, especially indoors. To support scalability, and eventually the potential for nationwide services, interoperability between these networks, and between the SAS systems, will be important. That is a central remit of the CBRS Alliance, which recently held an interoperability event at CableLabs’ facility in Louisville, Colorado, where over 55 different CBRS combinations were tested. The Alliance said there was a 98% completion rate across all these test scenarios, and no failures – an optimistic sign of a maturing ecosystem. GAA services are expected to start late this year, even while the FCC continues to mull over the final terms and conditions for the auctions in the PAL portion.
The stabilizing of a commercially ready ecosystem in GAA is highlighted by a rising number of FCC certifications for CBRS equipment. Ruckus Networks, now part of Arris, received the FCC’s first CBRS Device (CBSD) authorization earlier this month, for its Q710 and Q910 small cell products.
And last week, Ericsson said it had received this green light for its CBRS portfolio, which include indoor and outdoor small cells under its Ericsson Radio System banner. In CBRS, it offers the Radio 2208 for outdoor and a version of its Radio Dot indoor distributed radio system. The latter is significant, for those supporting new entrants and neutral host models in shared spectrum, because it is built for multi-operator deployment.
Of course, MNOs will use CBRS too, as they are looking to use 5 GHz spectrum for LTE-Unlicensed or LTE Licensed Assisted Access (LAA) in many parts of the world – adding affordable open spectrum to boost their overall airwaves. AT&T recently said it would use Samsung equipment for its initial CBRS fixed wireless deployment, harnessing CommScope’s SAS, and putting in place a migration plan to 5G when shared spectrum standards emerge. AT&T will test CBRS equipment in its labs early next year and plans commercial roll-out in certain US cities from late 2019.
This is an example of using current spectrum and technology for a particular application – fixed wireless – to complement both 5G FWA plans in mmWave bands, and the wireline business.
But some believe the use of shared spectrum will become less necessary in 5G, because there will be higher capacity bands available to MNOs, and early auction results, especially in millimeter wave, indicate their plentiful capacity will make them affordable. This is one factor that reduces the risk, to new open models, that 5G-Unlicensed will not appear for several years, although that also suggests that most of the innovation in flexible and neutral host platforms will take place, for a considerable time, in LTE enhancements. That is particularly likely, because there has been development of many specialized, vertically-focused offshoots of LTE standards recently, such as LTE-V for vehicles and NB-IoT for low power WANs, and it is unlikely that providers which invest in those now will want to change their technology for at least five years.
Of course, outside North America, those high frequency 5G bands include 3.5 GHz – which in most regions has been mainly used for fixed wireless but will now be a 5G band. Opponents of the flexible LTE framework of CBRS, notably T-Mobile USA, have argued that the FCC has put the USA out of step with the world, and restricted its 5G options, by not also making 3.5 GHz a 5G allocation).
Selected responses to FCC’s request for information on CBRS SAS plans:
- CommScope’s Comsearch business unit said it wants to demonstrate interoperability with a group of other SAS hopefuls – “at a minimum” that would include Amdocs, Google and Sony. Comsearch has also demonstrated interoperability with Ericsson’s CBRS kit.
- Key Bridge Wireless told the FCC it intended to conduct testing for 8-10 weeks to “produce a substantive and meaningful report for the Commission. Its initial commercial trials have been conducted with Nokia and it will host its SAS system on the AWS cloud.
- Amdocs aims to test its solution, CBRS Under Test (CUT), for a minimum of 30 consecutive days before completing its report for the FCC.
- Fairspectrum is a subsidiary of a Finnish company and has submitted a proposal for short term, limited geographic deployment in CBRS. It is not among the companies conditionally approved to be SAS administrators at this stage, but aims to apply to offer this service in the second wave.