Autonomous cars, 4K video, connecting the unconnected, augmented, mixed or virtual reality, tactile internet, industry 4.0 … the list of services and experiences that 5G could enable grows every day, writes Patrick Lopez, the vice president of Networks Innovation at Telefónica.
Will the evolution of technology be sufficient to enable all these use cases in one network? What are the necessary steps to ensure we build networks that will meet future expectations?
As telecoms network vendors increasingly ramp up their communication efforts towards 5G, our industry is once again struggling to frame a technological evolution in terms of customer benefits. While 2G was about voice, 3G about data and 4G about video, too often we saw the value propositions resumed to better coverage, more availability, more speed. While these factors are undoubtedly important, I don’t think that our customers care that much about 5×9’s availability or megabits per seconds. The problem with trying to iteratively improve our networks’ metrics is that it is too easy to lose sight of the customer.
Intuitively, anyone would agree that a larger, faster, more robust network is desirable, but practically, it is unlikely that any of these attributes would matter much to how our customers view our networks. In some cases, more capacity or speed can actually compound bad customer experience.
Take mobile video for example: With the ubiquitous implementation of technologies such as http streaming and adaptive bit rate, streaming video services have exploded. Conventional wisdom says that a bad user experience – long start time, bad picture quality, buffering – is mostly the product of congestion and would disappear if more capacity was applied. It is certainly an option in some cases, but in many, the problem is more difficult to resolve. For instance, if many people try to stream video from different services, the perceived quality or the users is more the result of negotiation and competition between the different video players than the product of network capacity.
Each device, video player and content provider will try to reserve as much capacity as possible from the network, immediately increasing their demand as more becomes available until the maximum video definition is displayed. The result is that video streaming service quality will depend very much on the device, and content provider. While this might be a positive for some phone manufacturers or content providers, it is not beneficial to the overall population. This is a clear example where focusing on increasing a network’s capacity yields little positive results.
So, how can 5G improve customer experience?
The traditional view of the industry can be summarised in two words: network slicing. Network slicing promises to give dedicated and elastic capacity ideally ensuring that each type of service will be able to independently scale up and down automatically. A large part of the industry views these slices as monolithic services such as voice, browsing, video, messaging, video calls … 5 to 10 slices should meet most needs.
The problem is that thinking only in terms of services is not very customer centric. Most of the industry has focused on selling one product: connectivity. With little variation, this product is the same for all, across all networks.
Differentiation in mature markets is difficult and price levels usually drop to a level where basic services are close to a utility model. In many cases, companies who want to generate more value will turn to personalisation. The connectivity market might not turn differently. The way you use voice calls, texts, the way you browse, the way you stream and produce pictures and videos in all likeliness differs greatly from your mother, your neighbor or your daughter. You are likely to use these services differently depending on the time of the day, whether you are at work, at home, on vacation. The possibilities are nearly infinite, but in aggregate, at the scale of a population, while these use cases are numerous, it is possible to use technologies such as big data and machine learning to detect patterns and isolate segments of usage, as well as segments of population.
These distinct segments can be harnessed to improve and create delightful customer experiences. There certainly are as many connectivity profiles as verticals, industries, type of enterprises. Consumer segmentation also shows marked differences in usage, expectations and type of services used by millennials, Gen X, road warriors, retired, independent workers.
I believe that the opportunities for network operators in 5G is to move from network slices to network strands to accommodate hundreds or thousands of user and services segments. The key is to create a network architecture that is open, programmable and elastic. Openness can be achieved with open source, open application programme interfaces (APIs) and standard interfaces, as well as a programmable orchestration model such as OSM. Programmability necessitates a software defined virtualized service environment as well as in-house developers that can translate market requirements into different flavors of existing services or new services. Elasticity of the underlying infrastructure is key and relies on a robust software-defined data centre, core network and radio architecture.
