Out for comment: The Citizens Broadband Radio Service for the 3550-3650 MHz band, featuring multiple priority levels, an annual auction, and an unprecedented Spectrum Access System.

As we have known for years, the FCC – and the rest of us – face a daunting problem: too many spectrum users and not enough spectrum. Even the most ambitious reallocation (and associated re-packing) plans offer at most limited and impermanent responses to the problem. Now the FCC has a new idea that could completely remake how the world handles spectrum and, in the process, dramatically increase the efficiency with which spectrum is used.

But the FCC’s plan to create a new “Citizens Broadband Radio Service” (CBRS) in the 3.5 GHz band sure has a lot of moving parts.

The plan is, in a way, simplicity itself: it calls for spectrum users to take turns. But while the basic concept may be kindergarten simple, its implementation is not, by a long shot. After all, not all users’ communications are equally urgent or important. Is there a way to get the more urgent and important stuff through first, and still give everybody else a fair shot?

After a year and a half, a previous Notice of Proposed Rulemaking, a follow-up public notice, hundreds of comments, two full-scale workshops, and too many meetings to count, the FCC has come up with a highly promising approach that is almost ready to test.

The CBRS would be established at 3550-3650 MHz, possibly combined with the adjacent 3650-3700 MHz band. This is an odd choice, at first glance. The incumbents at 3550-3650 MHz are high-powered ground and airborne military radars, and earth stations that receive satellite signals – on the face of it, poor candidates for sharing. The 3650-3700 MHz segment is widely used for delivery of commercial broadband service. Both bands have long been regarded as so sensitive that they are closed to most unlicensed devices.

The FCC nonetheless thinks it has a way to introduce new users to these frequencies without causing interference either to incumbents or to each other. The details appear in a Further Notice of Proposed Rulemaking (FNPRM).

Like an earlier plan, the proposal would establish three priority levels to govern access to the shared spectrum – but there the similarities mostly end.

The new plan would give the highest priority to Incumbent Access (IA) use. As the name suggests, IA would be limited to incumbent users. They would be permitted to operate at any time, and would be free to cause interference to anybody (except, in some cases, to each other). They would also enjoy protection against interference from everybody else, using geographic “exclusion zones” within which no one else could operate on the incumbents’ frequencies.

Next in the pecking order would Priority Access (PA) use. PA users would have to protect the incumbents, and would in turn be protected against interference from General Authorized Access users (discussed below). PA use would be auctioned in ways that are novel for the FCC. Auction areas would be small census tracts – about 74,000 of them nationwide, each with only a few thousand people, generally following political boundaries such as city lines. Each auctioned license would cover 10 MHz and be valid for only one year. Bidders would be free to aggregate licenses so as to cover a greater geographic area, or more bandwidth (up to 30 MHz) or more time (up to five consecutive years in a single auction). The FCC would conduct auctions for the entire system annually.

The lowest priority would be General Authorized Access (GAA). GAA use would provide access, free of charge, to spectrum not occupied by the two categories above. How much spectrum remains available for GAA will depend in part on the success of the PA auctions. But the FCC proposes limiting PA auctions to no more than 50% of the bandwidth in each census tract, after accounting for the incumbents, with the other 50% reserved for GAA. Still, because multiple users may compete for that 50%, no single user could be assured of getting a signal through quickly.

Certain “critical facilities” such as hospitals, public safety organizations, and local governments would get an advantage: they could reserve up to 20 MHz of GAA spectrum, but for indoor use only. Other GAA users would not be permitted to operate on those frequencies in those buildings. The FCC expects the building walls to largely isolate the users in these critical facilities from the great mass of other GAA users outside.

The key to making all of this work is a Spectrum Access System (SAS) that, as envisioned by the FCC, will manage all users (except the incumbents) on the fly, in real time. Every device will have to check in with the SAS, report its own location, request permission to transmit, and wait to be assigned a specific frequency. The SAS’s job will be to keep everybody off the incumbent spectrum in the exclusion zones, prioritize PA users, and assign GAA slots to others. Although a PA license would entitle the user to 10 MHz (or more) at any time, it would not specify a particular slot. For maximum flexibility, the SAS would be able to move each user’s actual operating frequency around in the band as needed.

The proposed technical rules are complex; those interested will have to plow through the FNPRM. Power limits will be higher in rural than in non-rural areas, and higher still for fixed systems, ranging overall from 1 watt up to 200 watts EIRP. But tighter controls will apply at the boundaries of a license area. A licensee that buys just one or a few census tracts will often be close to a boundary, and so in practice may have less power available. Presumably the SAS will set power so as to protect the boundaries.

Taken together, the technical demands on the SAS are unprecedented. The FCC describes it as an extension of the database that controls TV white space devices, but the tasks that would be assigned to the SAS would be vastly more complex. The design and operation of a successful SAS will be a major technological achievement.

All of these considerations listed here, plus a great many more, are open to comment.

Of the FCC’s earlier proposal for the 3550-3650 MHz band, we said, “Not content with crossing that tightrope, the FCC wants to juggle at the same time.” Now it wants to do the whole thing balancing tiptoe on a unicycle. But if the FCC and its industry participants can bring this off, and then extend the same ideas to other bands, the day when we finally exhaust the spectrum may not come as soon as we had feared.

The deadlines for comments have not yet been established. Check back here for updates. Anyone wishing to comment now may do so by uploading their submissions in Proceeding No. 12-354 at the FCC’s ECFS filing site.