Sharing Federal Spectrum by Contract
News reports led me to believe the administration had unveiled a national spectrum plan on October 25. Alas, the White House statement is simply a plan to make a plan, not an actual strategy. But the outlines suggest we’ll see progress over the current approach to sharing spectrum.
There will be a new spectrum strategy because there has to be one if our network-based economy is going to continue growing. Not only does the US need to deploy 5G at a speedier pace than the rest of the world, we need to keep feeding the public’s appetite for mobile apps.
The hard thing about developing strategies that serve the needs of today’s markets and technologies is finding an approach that won’t put itself out of business in short order. We’re focused on 5G right now, but much of the nation will be on 4G for quite a while; and some day, we’ll have 6G. We’re also not 100% sure how 5G is going to work. And we’re not even 50% sure about what apps will dominate 5G networks.
Revising the Current Plan is Crucial
Officially, the current spectrum strategy was developed by the Obama administration through such vehicles as the 2012 PCAST report on spectrum sharing. This report, effectively dictated by Google and Microsoft, emphasizes database-driven spectrum sharing systems along the lines of the (failed) TV White Spaces system.
The PCAST report was a disappointment because it failed to consider the only approach to increasing the utility and efficiency of spectrum-based systems that has ever worked, “upgrade and repack.” Google and Microsoft are both great companies with long histories of innovation, but creating novel ways to use spectrum isn’t really in their skill set.
The PCAST approach is essentially unworkable because it concentrates too much power in the keepers of the spectrum database (firms such as Google and Microsoft), too much coordination, and no capability to direct real-time behavior. It’s a classic Rube Goldberg contraption.
The Future of Spectrum Systems
Rather than building databases of possible networks location-by-location with complex, hierarchical permission systems controlled by government, engineers would rather build systems that naturally interact well with each other according to specific, objective, technical terms. The issues that need to be resolved are mainly about real-time coordination, and there are several ways to address them.
One approach for solving this problem was used in the Internet’s first radio-based system, the spread spectrum packet radio network deployed 42 years ago. Another system – Code-Division Multiple Access (CDMA) – applies sharing logic to process of converting bits into radio signals. LTE devices all have some form of CDMA.
Other methods include Space-Division Multiple Access (SDMA), Multi-User MIMO, and beam-forming. In years to come, we’re likely to see practical systems that use some form of angular momentum and even the phenomenon Einstein dubbed “spooky action at a distance“, quantum entanglement.
The leap from these technical developments to bureaucratically-defined spectrum access controllers goes in the wrong direction; it also doesn’t solve any real problems.
Reducing the Government Footprint
Segregating government systems that use spectrum from private ones only makes sense when the systems serve radically different purposes, and not always then. Some military systems jam spectrum commonly used for navigation or other purposes, for example.
If your application is all about rendering other systems unusable, you’re obviously not into sharing all the time. But you probably are OK with allowing civilian use when you’re not actively fighting the enemy, which is most of the time.
All it takes to enable this is system design that shifts to alternate frequencies when any band becomes unusable for any reason. Using a spectrum access system to warn users that warfare is about to commence is essentially tipping off the enemy, probably not a wise move.
Sharing by Contract
For all uses that simply move data packets around the most sensible forms of sharing are controlled by contracts: MVNO agreements, data roaming agreements, and FirstNet-style service plans that give premium access to safety-of-life applications and generic access to cat videos.
Contract terms that specify particular technologies can be changed as new technologies pass proof-of-concept. Laws and regulations specifying usage parameters serve a similar purpose, but are much harder to adapt to emerging technologies.
Every government agency that operates a spectrum-based system should be able to specify conditions under which it is willing to share with the private sector. Not all such terms are going to be reasonable at any given time, but each agency has a master; at least in theory if not always in practice.
The US needs a spectrum plan that allows us to transition from a hodgepodge of federal, private sector, and unlicensed systems to a coherent mixture of networks based on common technologies. There’s no reason for an agency to operate its own LTE network when there are so many competent operators.
There’s also no reason for an agency to continue using a pre-LTE system that offers no meaningful advantage over LTE. [Note: Substitute 5G for LTE if you’re reading this after 2020.] Hence, most agency spectrum use can probably be contracted out to the private sector.
Applications that can’t be supported by LTE and its progeny probably can be supported by a small number of alternative technologies that have commercial applications. So sharing by contract should be the default mode.
The kind of sharing that requires commercial fallback is a secondary mode, and there are other exceptions in which sharing by contract won’t work. The plan will need to specify conditions for these narrow cases.
Next time I’ll address the questions around broadband mapping and the technologies that aid it, LIDAR and HD Radar.