Alan Gatherer, CTO, Baseband SoC at Huawei and EiC of IEEE CTN
Neutral Hosting is an idea that has been kicking around for quite a while now. This month we take a brief look to see if its time has come. Here, by Neutral Hosting we mean the operator essentially rents the equipment running the Radio Access Network (RAN) from a neutral third party who "hosts" the software/algorithms the operator wants to run. The radio access is then hooked back into multiple operator networks to allow the operators to complete the network portion of the system. So why would anyone do this?
The driver behind this kind of Neutral Hosting is commercial enterprise; more specifically large venues such as hotels, stadiums, malls and the like. These businesses have customers and those customers want excellent wireless access. In fact wireless access has become part of the experience. Consider the 2016 super bowl, where Verizon reported 68 Terra Bytes of data over 9 days . But your favorite football stadium (by which I mean either American football or the real thing with the round ball) will also see massive amounts of data at every home game and such venues are actively integrating wireless access as an essential part of the entertainment experience. The problem for the venue owner is that getting great service for their AT&T customers is not a solution if they don't have equally good service for their Verizon customers and so on . So the significant effort the venue owner has to put into deploying equipment for one operator is multiplied by the number of operators their customers may subscribe to. The answer would seem to be have some sort of neutral system with only one equipment deployment, upgrade, maintenance and so on, but supporting all operators.
Also if there is one thing that the Over the Top (OTT) companies have taught us (specifically the group recently referred to as "technology's four horsemen of the apocalypse" ), is that there is money to be made in the ownership of data and especially metadata. The success of a venue may depend critically on understanding the customers in that venue and mining their meta-data. For instance, a vendor could be very interested in: the patterns of movement in a mall; relationships between customer movements and social media interaction; the use of online coupons; and so on. A stadium may have the opportunity to push related products or advertise related events that may be tailored depending on the online, real-time behavior of the customer.
A neutral hosted solution would change the competitive pressure on equipment manufacturers as well as operators. The parts of the system that could not be virtualized, for example the antenna and RF circuitry, would have to be shared by all operators. Operators could still differentiate themselves in areas such as MIMO and antenna density by simply "renting" more antennas from the host. However they could not differentiate with a superior pre-distortion algorithm for instance. The algorithms behind flexible antenna management are already becoming commercial, not for neutral hosting specifically, but for enterprise capacity improvement. See Airvana's onecell solution  as an example of this. On the other hand, the idea of flexibly renting equipment might allow operators to run a more efficient system by renting different antennas, perhaps even on an hour by hour basis, tuning their network capacity to suit their user density. There is an effect that we have called "the wave effect" that is the dual of the tidal effect seen in CRAN and Hoteling. This is where the amount of processing needed by a particular operator in a particular space is negatively correlated to the processing needed by another operator. More simply put, if all of the customers in a coffee line are Sprint customers, then there is no room in the line for AT&T customers and vice versa. Operators may be able to buy and sell computing capacity on the RAN in a very cloud-like manner, and over much shorter time scales based on predicting their customers' needs. Of course this kind of RAN trading is very futuristic, but it mirrors ideas in spectrum trading if we extend neutral hosting into the spectrum itself.
So where is neutral hosting today? Several companies are trying to sell some form or another of neutral hosting (for instance see [5,6,7]). There is also some effort in this direction in standards bodies with MOCN and MORAN . But these have not experienced much traction in the USA. 5G might provide an impetus for neutral hosting as 5G may rely on 4G to provide outdoor coverage and focus on indoor and dense area networks. As an emerging standard it can be developed with neutral hosting in mind. In 4G, outdoor and macro networks still dominate the revenue, making it hard to focus on such an unconventional market. But 5G may flip that equation. In addition, there needs to be an acceptable, at least de facto, standard interface to allow the operators to differentiate on shared hardware. The current hardware doesn't support shared access, as such isolation comes at a cost. But 5G may also flip that equation because of its focus on multi-use (MBB, uMTC, mMTC) modems. The industry is already discussing Service Oriented Radio (SOR) for 5G . So in summary, neutral hosting is certainly worth keeping an eye on. We are already creeping down this road but there is a long way to go and many business and technical hurdles need to be overcome before the progress really starts to accelerate.
Editor-in-Chief: Alan Gatherer (firstname.lastname@example.org)