5G will create wealth. The silver bullet to attract new industry segments is a guaranteed latency of no more than one millisecond.
5G’s frontrunners met in London this week to their share research results, thoughts and outlook on the technology’s low-latency appeal in attracting new vertical industry segments. The inaugural URLLC conference’s high-caliber speakers from mobile operators, infrastructure vendors, standardization bodies, and automotive/industrial/utilities/healthcare verticals revealed many opportunities and challenges to expect further down the road. In contrast to previous and current mobile networks (2G, 3G and 4G), fifth generation’s remit is to connect all kinds of appliances, machines, and things with each other – not only people. uRLCC’s low latency and high reliability stand to benefit the automotive industry in its quest for self-driving vehicles, factory automation where fast response times are crucial, and the so-called tactile Internet where humans control robots and machines in remote locations through touch combined with visual feedback.
uRLLC stands for Ultra-Reliable Low Latency Communications. It is a fundamental apex in the 5G triangle that neatly sums up what to expect from next-generation mobile (and fixed) networks. It is also the most challenging of the three.
Ultra-reliable and low latency communications is one of three 5G service categories as defined by the ITU. Source: ETSI presentation - Adrian Scrase - URLLC 2017, London, UK, 14 November 2017
uRLLC needs to resolve two conflicting requirements at the same time: low latency and ultra-high reliability. Guaranteeing low latency requires packets that are as short as possible, whereas reliability means adding extra bits to shield the channel against errors, or by simply retransmitting lost packets. For 5G uRLLC reliability equates to the probability of a failed packet delivery being less than 10-9. In other words the correct data is received 99.999999999 % of the time.
Latency sensitivity depends on the use case
“One millisecond” is the answer most in-the-knows provide to define 5G’s latency. Yet that’s only a partial truth. It’s all about achieving realistic reliability and latency parameters. As Adrian Scrase, CTO of the standards body ETSI points out, low latency requirements are very much application dependent.
"The latency demand changes by use case increasing as you move up from the URLLC to eMBB apex on the 5G triangle" - Adrian Scrase, CTO of ETSI
In simplest terms, latency is the total time it takes for data leaving a source to arrive at its destination (one-way time) as it traverses through the network’s access, distribution, and backbone layers. Shortening the path is a key tenet for keeping the delay at a minimum.
Distributing intelligence in the network
Decomposing the upcoming 5G network into its enabling technologies and components testifies to its immense flexibility: a single physical network that is dynamically programmed to deliver not only uRLLC, but also mMTC (massive machine-type communications) and eMBB (enhanced mobile broadband) services at the same time. The network uses the efficiencies and computing/storage scalability of data centers – the cloud – at multiple levels: near the edge, at the aggregation level, near the core. Software-defined networking (SDN) and network function virtualization (NFV) provision a dynamically-configurable, service-oriented architecture. This centralized approach is very different from most existing telecom networks that use specialized hardware components distributed geographically.
With this kind of programmability, it is crucial for service requests that demand real-time processing and storage to be in close proximity to the requesting device. In those use cases with extreme latency sensitivity this means processing and caching occurs at the radio access level (base station), or at least very close to it.
"The distribution of intelligence in the network will be key to meeting latency requirements" - Pierpaolo Marchese, Standard coordination Head from Telecom Italia
To get a clear picture of what’s possible in reality, British Telecom analyzed traffic and associated latencies on 8000 cell sites in their current live network. By modeling results to potential 5G scenarios, BT found that latencies vary greatly depending on the functional split of the network. Even if the one-millisecond parameter might still be theoretical, the demands of latency-critical 5G applications requiring less than 10 milliseconds are most likely to be met.
"For 5G uRLLC it will be extremely important to break down and distinguish applications into three separate categories: those requiring low-latency, those demanding ultra reliable transmission, and those requiring both." - Andy Sutton, BT's chief network architect
Network slicing and net neutrality
For 5G, current networks will evolve into one single network encompassing both fixed and wireless domains. NFV will enable configuring the network on-the-fly to deliver the performance required by the service type. These so-called network slices allow addressing eMBB, uRLLC, and mMTC use cases in parallel on the same physical network.
However, network slicing stands in conflict with the concept of net neutrality on which the Internet was originally built: treat all data and traffic equally with the same priority. As such, regulators in different regions of the world are pursuing different policies regarding network neutrality. For example, net neutrality in China and Europe requires equipment vendors to take different considerations into account, as pointed out by Xueli An of Huawei.
"EU regulators should work together with the telecom industry to define net neutrality." - Xueli An, principal researcher at Huawei Technologies
It’s not hard to foresee that net neutrality impacts network slicing. Countries and regions with stronger net neutrality regulation could be at a disadvantage here, and this might ultimately burden the speed of 5G rollout and service uptake in those territories.
The European 5G challenge
In Europe 28 member states need to find a consensus on 5G. That’s far more difficult than for single countries such as the USA, Korea, Japan and China. It will also take more time. To ensure that Europe remains competitive in 5G on a worldwide scale and regains its former leadership in mobile technology, a joint initiative called the 5G Public Private Partnership (5G PPP) brings together the European Commission (the public side) with telecoms equipment manufacturers, mobile operators, research institutions and technology start-ups (the private side). 5G PPP is funded by the European Union (EU) with a budget of € 700 million for 5G research, development and innovation. After completing 19 initial R&D projects, 5G PPP began its second round this year with a further 21 projects. The 5G Infrastructure Association (5G IA) represents the private side of 5G PPP and its chairman, Colin Willcock, notes that despite the ambitious task of coordinating and agreeing on 5G efforts across so many countries, there are clear advantages of scale owing to the sheer size of the European market.
„In 2020 the UEFA EURO football championships will showcase 5G technology across Europe.“ - Colin Willcock, chairman of the 5G Infrastructure Association
5G PPP projects explore key 5G research themes and their R&D results serve as strands for standardization. The goal is to ensure a strong contribution of European know-how to the 5G standardization process.
Guaranteeing latency and reliability
Once 5G networks are in operation, their use by vertical industry will largely depend on meeting defined latency or reliability specifications. Given the many permutations on the sliding scale of latency and reliability, how do you assure industry verticals that the 5G network fulfills its promise? Meeting the demanding metrics will require contracts between the service provider (network operator) and customer (industry vertical) on quality, availability, and parameters critical for the use case.
"Service-level contracts - SLAs - between the service provider and customer will play an essential role in 5G ." - Anthony Magee, business development director at ADVA Optical Networking
Using network slices as a fundamental line of demarcation, service-level agreements (SLAs) for 5G uRLLC will capture all the critical metrics needed to guarantee latency and reliability.
The 3rd Generation Partnership Project (3GPP) unites seven international telecommunications standard development organizations responsible for fixing the specifications that define 5G. Their next release 15 is due out in the middle of 2018. It will freeze details of 5G’s New Radio (NR). As the name suggests, this new air interface includes radio access technology that is no longer backward compatible with existing cellular standards in order to satisfy uRLLC’s demanding latency and reliability requirements with new packet, frame and multiplexing techniques at the physical layer.
5G’s uRLLC will further evolve as 3GPP turns its attention to the subsequent release 16. Mark your calendars for next year’s two-day URLLC 2018 conference on November 13-14 in London when the experts meet to discuss the latest developments.