- C-RAN’s adoption is likely to grow significantly soon, thanks in part to evolutions in the underlying technologies.
- Long term, future RANs will see a dynamic mix of centralized and distributed functions.
In 2016, we’re likely to hear even more about C-RAN than we already have. It’s not a new concept, and plenty of operators have deployed mobile access network architectures in which the baseband processing units are centralized, stacked or pooled, linked to remote radio units elsewhere. As portions of the network become increasingly virtualized, baseband processing will become virtualized, too – thus, centralized RAN will evolve into cloud RAN. This won’t happen everywhere, of course, but its use is likely to spread thanks in part to some significant advancements in C-RAN technology coming soon.
Some RAN vendors have already promoted plans to replace the notoriously inefficient CPRI protocol interface between baseband and radio with Ethernet links, making it easier for operators to use copper or microwave media for fronthaul in addition to fiber. With that fiber restriction dropped, it will become more cost-effective for operators to deploy C-RAN even where they don’t have fiber and can’t afford to deploy it. Thus, C-RAN’s reach is likely to grow.
Where does this trend go if extrapolated based on its apparent current trajectory? We occasionally hear the prediction that mobile access networks will ultimately evolve to the point where they are essentially boiled down to a data center on one end of a fiber and an antenna on the other end.
But, we can already see what’s wrong with that picture. All the functions of the mobile access network won’t simply be sucked into the data center like a black hole. In some cases, the opposite will happen.
Some vendors are already marketing mobile base stations that include integrated servers to do localized application hosting and processing and, potentially, content delivery. The value of these solutions is precisely because they offer distributed, rather than centralized, functions (increasing responsiveness to local needs, reducing transport bandwidth requirements, etc.). And they could have a major role to play in enterprise networks in particular, a key growth target for vendors and operators alike. So, as baseband processing gets sucked into the cloud, other functions will be pushed out to the edge.
In addition, it’s important to remember the significance of keeping network latency low going forward. Ultra-low latency is a key requirement of 5G networks and a key challenge for vendors and operators. As operators weigh the benefits of cloud-ifying portions of their baseband units, they’ll also be weighing the potential impact on latency of those decisions, which may be a gating factor impacting how widely cloud RAN is deployed.
If not the data center and the antenna, what will the future RAN look like? In dense areas, we’re likely to see functionality allocated among a number of centralized and distributed resources (data centers, “edge clouds,” base stations, enterprise locations, aggregation points, hetnet coordination points) with the ability to dynamically shift from one location to another as needed. It will be one network during rush hour and another late at night, one network on 14th Street and another on 138th. It will be highly coordinated and hopefully efficient. And it will probably be unlike what was predicted in 2015.