There are several problems to overcome in making spectrum sharing a reality. Some are technical but more often than not, it is easier to overcome those hurdles than commercial ones. So what are those challenges? And when can we expect to resolve them to make spectrum sharing a reality?
The technical issues revolve on managing interference. Different solutions are proposed to address interference. Cognitive radio, small cells and location-based access to spectrum are fundamental technologies. While certain, very basic, elements of these technologies have been implemented today, much more development is required to make a full spectrum sharing scheme practical. For example, to access TV white space frequencies, devices check a geolocation database for available TV channels prior to operation. However, the database does not provide any information on what systems already operate in that channel. Another example is Dynamic Frequency Selection (DFS) which is an operational requirement for systems in the unlicensed 5 GHz band in many countries. The idea is to have the device detect radar signals and vacate the channel if those signals exceed a certain threshold.
Detecting interference is a seemingly simple problem that has proven to be highly non-trivial. Specifically, detection of low power signals is challenging (not to mention from devices that are mobile or temporarily inactive!). There was initially a plan to mandate interference sensing for TV white space operation but the idea was not pursued due to technical difficulties. Today, much on-going research in cognitive radio technology focuses on developing receivers capable of detecting low power signals.
But managing interference is not just simply one of detection. Frequency coordination and access by different types of systems is another challenging aspect. Coordination and access can be manages and separated on a number of orthogonal planes: time, location, frequency, code, and other planes. To enable this, a number of technologies, processes and regulated frameworks need to be established. For example, PCAST’s recommendation to make 1,000 MHz available for shared spectrum is correctly founded on the premise that a large allocation of shared spectrum provides higher efficiency. Wideband transmitters and receivers would be advantageous in addition to technologies like carrier aggregation and channel bonding. Unfortunately, wide bandwidth increases the challenge on cognitive radio to detect interference.
For Federal spectrum, users are classified according to a three-tier hierarchy as primary (federal), secondary and tertiary users with different rights and privileges. Service interruptions and delays to secondary users are bound to happen when primary users come online. Policy control and coordination between secondary users is in itself another issue to be tackled. This point towards different shades of spectrum sharing arrangements that needs to be developed. Moreover, when the primary user is the government, oversight is bound to be more complicated than between like services.
Standardization, certification and compliance to regulatory requirements are another phase that has to be surpassed before deployments. Considering all of the above challenges, it will be sometime, as much as 10 years if not more, before we see dynamic spectrum sharing implemented. But more importantly perhaps is the question of how dynamic the spectrum sharing implementation will be.