Resolving the 6 GHz Conundrum
The FCC’s current proceeding on the 6 GHz band is a turning point in US spectrum policy. The Commission proposes to allocate 1,200 MHz of spectrum for Wi-Fi and potentially other unlicensed services. This would be the biggest private sector spectrum giveaway below the mmWave bands in US history.
It’s a very complicated proceeding because the band in question – 5.925-7.125 GHz – is already used by multiple services such as fixed point-to-point (common carrier) services, Fixed-Satellite Service, Broadcast Auxiliary Service, and Cable Television Relay Service. Rather than terminating or relocating the leases held by incumbent services, the FCC proposes to protect them from Wi-Fi interference through a patchwork of automated frequency coordination (AFC) systems and power limits in four distinct sub-band groupings.
The agency believes its technical analysis – and a report compiled for the (Silicon Valley) supporters of the plan by RKF Engineering Solutions – puts the plan on solid ground, but current license holders aren’t so sure. Electric utilities, public safety, and railroads are all pushing back, unconvinced that the plan includes sufficient means to both guard against interference and to mitigate it when it happens, which it surely will.
Arguments in Favor are Largely Generic
Arguments in favor of the plan are mainly generic praise for Wi-Fi and the Internet, largely ignoring alternative spectrum bands that could be, and in some cases already are, used to expand Wi-Fi’s spectrum footprint. The proponents pay lip service to 5G, but fail to consider the question of allocating at least some of this 1,200 MHz to licensed services.
The RLAN Group, which includes Apple, Broadcom, Cisco, Facebook, Google, Hewlett Packard Enterprise, Intel, Marvell, Microsoft, Qualcomm, and Arris/Ruckus Networks, argues for an expeditious proceeding without mentioning the difficulty unlicensed spectrum has with solving the 5G spectrum crunch:
Because access to this spectrum is so critical, both to meet growing consumer demand for Wi-Fi and to support other 5G investments, we ask that the Commission move quickly to resolve this proceeding and adopt rules that allow for rapid product deployment to maximize the value of the 5 GHz band for the country.
It certainly is the case that Wi-Fi needs more and better spectrum than it currently has. The 2.4 GHz band is only good for low-bandwidth services because its allocation is limited to 75 MHz. The 5 GHz band is power limited and constrained to only six 80 MHz channels or two of the ultra-wide 160 MHz channels that 802.11ax wants to use.
But the FCC and its petitioners have failed to show that 6 GHz is the best of all possible bands that might meet their needs.
Wi-Fi Suffers from mmWave Amnesia
When 5G proponents came to the FCC for spectrum to meet their needs, the agency kicked them up to the mmWave band, above 24 GHz. It also promised to free up 200 MHz of mid-band spectrum in the C-Band, but that’s still a work in progress.
There is a lot of skepticism about the utility of mmWave because it has issues with foliage and water vapor, but I’m convinced it will work as long as our receivers can aggregate reflections with direct signals. My belief isn’t stopping market analysts like Walter Piecyk from expressing doubts.
Neither the NPRM nor the RLAN people mention IEEE 802.11ad and 802.11ay, the standards that support Wi-Fi with over 2,000 MHz of spectrum around 60 GHz. 802.11ad was released in 2012 and the 802.11ay update is expected this year. This standard, commonly known as Wi-Gig, originally supported a reliable 5 Gbps and the update is expected to support 200-400 Gbps. That should be sufficient to meet most indoor needs for some time to come.
The Wi-Gig Alliance included most of the RLAN companies as members, by the way.
Wi-Gig has Moved Slowly
You can buy Wi-Gig routers and adapters today from Netgear and several other manufacturers. The Netgear Nighthawk X10 router line has it, along with some really nice features such as a 10 Gbps Ethernet port. Adapters can be had for $22, but there’s not much demand for them despite the fact that 60 GHz isn’t bothered by water vapor.
Not only do very few people need 5 Gbps in their homes and offices, consumers are put off by the fact that Wi-Gig doesn’t penetrate walls. The latter issue is often cited as a show stopper, but it’s an easy engineering challenge to solve. It would not be difficult to build a device consisting of two antennas, 4 inches of wire, and a Wi-Gig repeater that would solve the wall problem.
Drill an eighth inch hole through the sheet rock dividing two rooms, stick a matchbox-sized device on each wall, connect them with two wires, provide power, and you’re done. You could even do this in a rental as long as you fill the tiny holes with spackling compound when you move out.
The Wi-Gig industry hasn’t devised such a system because the demand is not there for it, at least not yet. [Incidentally, the “can’t penetrate walls” thing was raised as an early objection to 5 GHz Wi-Fi, but we got over it.] Nevertheless, Wi-Gig enjoys the same status today as Ultra-Wideband, a system proposed by IEEE 802.15.3a and the Wi-Media Alliance during the Mike Powell FCC days.
I Expect People will Use 6 GHz Wi-Fi – Some Day
People found 802.11ac good enough for video streaming so Wi-Gig wasn’t necessary. Similarly, they found Bluetooth 4LE a great upgrade from traditional Bluetooth, so there was no market for Wi-Media. So what would the fate of 6 GHz Wi-Fi be?
Even though the business case for a new spectrum band for Wi-Fi 6 looks strong, history shows that Wi-Fi upgrades are very slow to materialize. Wi-Fi 6 isn’t yet supported in many enterprise Wi-Fi product lines, such as the popular Ubiquiti Networks Unifi system. The company does offer a gamer-oriented system called AmpliFi, but its enterprise access points still use 802.11ac.
Using a Wi-Fi upgrade means, for most people, buying new computers (and phones and tablets…) as well as installing new routers and access points. This process is slow, so we won’t see significant penetration of Wi-Fi 6 for 3 – 5 years. Without Wi-Fi 6, there’s not much point in assigning 1,200 MHz of new spectrum this year.
We Have the Time to Do This Right
I can certainly appreciate the desire of public servants to make bold moves in the public interest and the difficulty everyone in Washington has with making prudent decisions in this time of political polarization. But these factors don’t justify making such a huge investment in Wi-Fi when so many interference issues are unresolved.
Assigning unlicensed spectrum is a much more consequential move than selling licenses because once a Wi-Fi device is certified no one is accountable for the problems it may cause in the field. The FCC can’t come into your office and make you point your antennas in a different direction, it can only instruct vendors to issue software updates that turn off access to particular channels. And it almost never does that.
If the FCC scales back its ambitions for the near term, we can gain valuable experience about what happens in the real world as Wi-Fi 6 attempts to share spectrum with incumbent licensed services in the 6 GHz band. It would be reasonable to release enough spectrum this year to support an additional three, non-divisible, 160 MHz channels for ultra-high-speed use.
Learn and Scale Up
As former FCC Chief Engineer Julie Knapp likes to say, the mantra for spectrum re-allocation to support new services used to be “upgrade and repack.” We upgraded our old analog TV system to digital for higher definition and better overall quality and found we could do a better job with less spectrum.
So we had a digital dividend of unused spectrum that we could spend in 4G licenses, TV White Spaces, and other things. This doesn’t happen with Wi-Fi because it has an insatiable appetite for new spectrum and zero incentive for more efficient use.
We need to learn how to mitigate the interference Wi-Fi 6 is going to cause in the 6 GHz band before going whole hog with 1,200 MHz of new spectrum. This means license holders and the FCC must monitor and report instances of interference as well as whether and how they’re resolved. This kind of information will allow us to validate and refine the predictions made by the engineering models.
The unlicensed mantra should be “learn and scale up.” Unlicensed is the only area where regulators still use the “beauty contest” method for allocating spectrum, a mechanism best characterized as “a rent-seeking process that made socially wasteful investments in pursuit of political favors.” Hence, they need to base each decision on rigorous data rather than on generic hand-waving about the wonder of it all.
Wi-Fi Needs to Get Better at Coordinating with Wi-Fi
While the NPRM’s interference analysis emphasizes interference between Wi-Fi and non-Wi-Fi systems, the fundamental dilemma is the way Wi-Fi networks treat each other. In principle, a 160 MHz channel using Wi-Fi 6 with 1024 QAM and 980 sub-carriers can teach a top throughput of 9.6 Gbps.
It stands to reason that a properly shared 9.6 Gbps channel should be able to meet the needs of the handful of access points that can all see each other on at any given time. We only need a massive number of channels if access points are not coordinating with each other efficiently. Enterprise systems that coordinate well are able to work well on a single channel.
So the challenge for Wi-Fi is to take up the mantle of inter-access point coordination in a serious way. This has always been a problem that 802.11 has swept under the rug by pleading “privacy” or some other nonsense.
Wi-Fi/Wi-Fi Sharing is the Key Problem
Yes, inter-access point coordination is not an easy problem to solve; but it’s an important one. It’s necessary to constrain Wi-Fi’s bandwidth appetite because it’s not the only system in town.
If the Wi-Fi industry can roll out an upgrade that works efficiently over three contiguous 160 MHz channels and still needs more spectrum, that’s a conversation we can have five years from now. Here are some hints for how to solve the problem:
- Reserve the new channels for devices that meet Wi-Fi 6 specs for 160 MHz channels.
- Run all devices at the fastest practical data rates in order to free up air time for each other.
- Assign channels so that neighbors group on common channels to the extent practical.
- Employ a meta-group protocol so that users form cells running common channels adjoining other groups using different channels.
- In other words, operate Wi-Fi more like a cellular network and less like a riot.
Wired enterprise networks manage to share gigabit Ethernet switches with hundreds of users without incident. There’s no good reason that we can’t expect this kind of behavior from Wi-Fi 7.
Run This as a Pilot Project
I’m proposing that the FCC releases 480 MHz of bandwidth in the 6 GHz band for a pilot project. The terms of the pilot are as specified, three high speed, indivisible 160 MHz channels supported by ongoing work on inter-access point coordination.
I’m also open to experimentation on the MAC protocol, which is far from perfect. And I’m also proposing a rigorous monitoring of interference issues.
If this goes well, we can explore any additional allocations that may be necessary in either the 6 GHz band, the 7 GHz band, or in the mmWave bands. It’s way too early to give up on Wi-Gig because the mmWave territory is perfect for networks confined to individual offices, homes, and apartments.
The FCC should also auction some licenses in the 6 and 7 GHz bands just in case the mmWave band doesn’t work any better for 5G than it has for Wi-Fi. All in all, we need the agency to continue playing fair with the mature technical wizards of the 3GPP world and the rogue geniuses of unlicensed.