Apple will soon join the 5G ranks with Samsung, LG, OnePlus, Motorola and others when it integrates 5G into its next generation iPhone in 2020. And with CES 2020 and Mobile World Congress right around the corner, there is no doubt that 2020 will prove to be yet another action-packed year of 5G handset launches. Despite the initial buzz we’re seeing in the media, however, most carriers are still struggling with the 5G business case. But one thing is clear, it won’t be driven by mobile broadband consumers.
Early indications are that we’re starting to see from the early 5G networks that they are extremely fast, while the quality of 5G coverage is inconsistent at best. Eventually, of course, carriers will iron out the early wrinkles, but will consumers be willing to pay more? Not really. Sure, some early adopters will be willing to pay more, but the vast majority feel consumers will only pay 5% more for 5G than they currently pay for a 4G package.
In fact, many operators are still struggling with pricing strategies when it comes to 5G. In the U.K., British Telecom (BT) concedes that charging a premium for 5G could be unsustainable as competitors like Three have signaled they’re willing to offer unlimited 5G at no extra costs. Verizon, on the other hand, was blamed for ‘sneaking in’ a $10 premium for their 5G over 4G plans. No matter how you slice it, many in the industry are starting to wonder how operators plan on recouping their investments in 5G and turn a profit.
There is still a business case for 5G
Over the last few years, a slew of ‘new’ 5G services promised multiple ways for operators to generate new revenue streams, yet most are far off at best. For example, autonomous vehicles: while captivating, the reality of autonomous vehicles is still years away, requiring massive investments in R&D and the jury is still out on who actually benefits from the revenues. Likewise, massive machine communications, driven by the internet of things (IoT), has gained incredible momentum. Yet most will agree that the current needs of IoT can be addressed with current LTE infrastructure and Wi-Fi. Most consumers can easily roll these extra bandwidth requirements into their current unlimited plan.
There is one business case left which has not been displaced - real time video. Real-time video requires ultra-reliable low-latency communication (uRLLC), one of the several different types of use cases supported by 5G. There are two prime examples of real-time video in action - cloud gaming and remote surgery. Both are currently being served by dedicated connections or local implementations, often at exorbitant rates.
That said, providing the assured, real-time capabilities required is not simple. Transitioning your transport network to accommodate network slicing will help. But current 4G radio and transport networks cannot guarantee the SLAs required by URLLC.
Now we’re talking about slicing
Real-time video requires unique requirements, SLAs and profiles, and all this will require network slicing. In a world where network resources are limited, network slicing is the technology of choice for network architects and industry bodies (IETF, ITP, and others). Yet, the same industry bodies differentiate between different types of slicing - hard and soft.
According to the IETF, hard slicing refers to the provisioning of resources in such a way that they are dedicated to a specific network slicing instance. Soft slicing refers to the provision of resources in such a way that the slices are separated so that they do not interfere, on average, with each other. However, they can interact dynamically, which means they may compete for some resource at some specific time. In other words, ‘soft slicing’ promises delivering on SLAs on average, while ‘hard slicing’ dedicates fixed resources to specific services or customers.
To make things even more complicated, there are a variety of technologies which can be used for achieving hard and soft slicing. For some of you on the packet side of the house, soft slicing is a mere extension of statistical multiplexing paradigms. And for the most part, statistical multiplexing is very efficient when there are no QoS guarantees or when they are soft. Trouble arises when we need to guarantee a high degree of precision.
On the other hand, today’s network elements are not capable of hard slicing. This means that to incorporate hard slicing, most carriers will need to upgrade their networks - in many cases ripping and replacing network elements. Moreover, the inability to ‘over-provision’ inherent to hard slicing technologies, puts a crimp in the potential revenue stream of carriers.
It is for this reason that slicing capabilities need to be dynamic and intermittent in nature. To enable the sale and resale of the same resources, on the fly (dynamically) and for limited periods of time. Even the IETF expects that the frequency of reprovisioning with network slicing will be relatively high. The limited nature of uRLLC services (i.e. for the length of a game/match or a surgery) makes this reasonable.
Network slicing cost vs. benefits
Talk to most carriers exploring network slicing and they’ll tell you the benefits of network slicing are not clear and that they don’t outweigh the complexities and costs of re-architecting the network. However, network slicing is currently the only way to efficiently and cost effectively deal with multiple services types. The upside? Network slicing will improve the chances of ROI, by profiling services through SLAs and offering differential pricing. I believe that over time, all types of slicing will be used, both soft and hard.
Since rolling out network slicing, is no different than incorporating new technologies in your network, you know how time consuming and complex the process is. Slicing has many flavors and many technologies, but currently only hard slicing, which dedicates fixed resources to specific services or customers can assure uRLLC services, such as real-time video. And enabling dynamic, hard slicing, can secure future revenue streams which will help you recoup your investment.
