A new shared spectrum scheme is being introduced in the US at 3.5GHz called CBRS, ideally suited for in-building small cells. It promises to unblock the logjam by opening up new spectrum for almost anyone to use with standard LTE on future mainstream smartphones. What exactly is being proposed, how will it work and where else could this approach be used?
What is CBRS?
Don’t confuse this with Citizen’s Band Radio, a two-way voice system at 27MHz most commonly used by long distance truck drivers.
CBRS (Citizens Broadband Radio System) uses TD-LTE to provide a wireless voice and data service at 3.5GHz (3550MHz to 3700MHz). The radio interface is exactly the same as for LTE at other frequencies, supporting voice, text and data services with seamless mobility. This frequency allocation spans existing LTE Band 42 (3400MHz to 3600MHz) and LTE Band 43 (3600 to 3800 MHz).
What’s different is the way that the spectrum is assigned to each user. This isn’t sold to operators in large blocks covering wide geographic areas nor a completely unlicensed free-for-all (such as Wi-Fi). Instead, use within each building is individually requested and assigned on a case-by-case basis. Where it is no longer required, it is returned for use by others.
The process of assigning spectrum is automated, with several Spectrum Allocation Servers (SAS) co-ordinating the scheme nationwide. As part of the registration request, each Small Cell reports its position to within 50m horizontally and 3m vertically. The SAS uses terrain data and radio propagation models to calculate the impact for other nearby small cells. Additional outdoor radio measurement receivers (Environmental Sensors) are also used to assess background levels. If the RF power density is less than -80dBm (1x10-11 Watts), then the SAS authorises spectrum use.
The philosophy is that spectrum is a precious national asset and should be used rather than lie dormant.
FCC Rules Part 96 specify the system operation in great detail.
The scheme defines three levels of priority when assigning shared spectrum:
- Incumbent users. Specifically US Naval Radar operates in this band today, and would have first priority. This is more applicable in the coastal areas but can stretch quite some distance inland and does cover many of the more populated areas of the US. There are other existing users of this band who retain grandfather rights. However, the 3.5GHz band doesn’t tend to penetrate indoors as much as lower frequencies, which should allow many modern buildings to reuse it – those places with poor inbuilding cellular and which have strongest need for CBRS are most likely to be able to be assigned spectrum.
- Priority Access: An organisation can pay a fee to be assigned a 10MHz block for a limited geographic area (such as a town or district) for a three year period, giving them priority over other users. No more than seven such blocks would be concurrently allocated within the same geographic area (i.e. 70MHz out of 150MHz available).
- General Authorised Access: The standard service used by the majority of users.
The fee structure and pricing levels charged by SAS providers are as yet unknown, but required to be "reasonable" and could be regulated if found not to be so.
How would this work commercially?
Not too dissimilar to installing Wi-Fi, building owners would be able to buy and deploy their own CBRS Small Cells which would automatically register with a SAS. The SAS providers will charge a small fee for spectrum allocation – I’m estimating hundreds of dollars per installation per year, but the amounts aren’t yet known.
It is likely that professional Wi-Fi installers and in-building wireless experts will be in demand for anything but the simplest/smallest buildings. These companies would physically install, commission and upgrade onsite equipment as and when required.
Such systems could operate in a completely standalone mode as a private network. More commonly, roaming with the major cellular networks will be required. This is likely to be handled by a relatively small number of “neutral hosts”, who will aggregate and consolidate traffic from many smaller systems and feed through to the major networks.
The CBRS Alliance was formed in 2016 by interested companies to promote the technology and drive adoption.
The radio interface is standard TD-LTE and so should be relatively easy for smartphone vendors to incorporate – it’s yet another frequency band, so not trivial, but is more related to antenna design and the RF front end.
3.5GHz LTE is already in use in other parts of the world, including Japan.
I would think it reasonable to expect commercial products to be available during 2H 2017, but whether this would make it in to the iPhone in 2017 is anybody’s guess. My forecast would be for September 2018, subject to commercial take-up and operator co-operation.
Small Cell products
We’re starting to see the first commercial CBRS small cell products appearing. The vendors have had to develop and test the back-end interworking with SAS databases, the first of which have now been completed. Field trials are likely during 1H 2017 with full commercial service available in the second half.
Specific announcements include:
- SpiderCloud completed product development of dual-band radio nodes including interoperability testing with Federated Wireless SAS. They will start field trials in Q1 2017.
- Ruckus Wireless demonstrated end-to-end at Dell EMC event on October 17. They brand their CBRS Solution as OpenG.
- Nokia announce CBRS 3.5GHz band support for their Flexi Zone outdoor microcell and indoor picocell.
- Several other LTE-only small cell vendors have plans and could move quickly where there is market demand. Press releases announce intent from Telrad, Airspan, Accelleran, BaiCells, Ericsson, ZTE and more.
Professionally installed small cells can have their location manually registered by the installer. Portable "non-professionally installed" small cells must be capable of reporting their location automatically, including if and when moved. This would require at least onboard GPS and possibly better location tracking methods when installed deep inside buildings without sight of the sky.
There will also be a requirement for neutral hosts with LTE small cell gateways, LTE core networks which interwork with both fixed and mobile networks. Such products already exist and progress will be more related to commercial investment and operator roaming agreements.
Spectrum Allocation Server (SAS) and Neutral Hosts
Federated Wireless, an independent start-up in 2012 and now subsidiary of Allied Minds, is purely focussed on SAS and perhaps the most vocal and visible. Several large companies are also developing their own, notably Google and Nokia. A handful of others have been mentioned, but there are few public announcements outside of the above.
SAS systems are now quite mature and have completed interoperability testing with several vendors. I’d expect Federated Wireless to become approved for operation by the FCC before the end of the year.
Google are also thought to be considering providing a neutral host gateway service to simplify connection into the major networks but I can find no public details. Other neutral host gateways could be used - these functions are quite independent - and may include some core network functionality. MVNOs with their own core networks may have an opportunity to aggregate/interwork CBRS installations.
Shared Spectrum in other countries
The concept of shared spectrum is capturing interest worldwide. Many regulators are viewing this US initiative with a view to adopting a similar scheme if it proves successful. It may also form a foundation for 5G New Radio spectrum allocation or even for use in (small cell) wireless backhaul microwave.
The UK regulator Ofcom published a statement on Shared Spectrum policy in April 2016, and is considering adopting this type of scheme for the 3.8 to 4.2GHz bands.
Shared Spectrum is also being considered by the Telecom Infrastructure Project (TIP) project to expand telecommunications provision for developing countries and remote areas.
The calculation of spectrum assignments and power levels is critical to ensure high spectrum reuse and efficient allocation. One idea (from iPosi) is to use a wider range of in-building measurements to assess the RF isolation of the building (i.e. how much of the small cell signal would leak outside).
I would expect most users to be either connected to the CBRS or directly to their normal network operator at any one time. There is potential to achieve high data speeds by aggregating channels across both using Carrier Aggregation. This would require tight co-ordination between the CBRS and cellular operator – probably directly connecting the CBRS radio nodes into the operator’s core network – and also phase synchronisation throughout. In order to simplify the huge complexity of channel aggregation, 3GPP standards define a limited set of channels which can be combined and at the moment only support 3.5GHz (Band 41/42) with the 1800MHz band.
I think we will see some momentum growing for CBRS during 2017, both in small cell products and supporting devices.
What will be more critical is the establishment of independent neutral hosts and approved installers. Support (particularly via roaming) from mainstream network operators will be critical.
Depending on the success of this approach with CBRS, we might expect to see the scheme adopted elsewhere and in other frequency bands in the next 3-5 years.
Pros and Cons
|Phones likely to be more easily/quickly available||Needs capable smartphones|
|Uses full capabilities of LTE unlike MulteFire||VoLTE essential for voice|
|Co-ordinated between small cells (unlike Wi-Fi)||Limited potential for Carrier Aggregation for highest speeds|
|Independent of exclusively licensed frequencies for individual operators||Unknown pricing/fees|
|Ideal for neutral host/shared infrastructure|
|No huge spectrum fees|
|Makes better use of available spectrum |
(use it or lose it)