The third wave of the 3GPP Release 15 standards – the so-called ‘late drop’ – has been released, several months later than originally expected. They were ratified at a recent meeting of the 3GPP RAN group in Shenzhen, China. This was not accompanied by much fanfare – it mainly consists of highly detailed technical specs, many concerned with dual connectivity (see separate item). However, the existence of multiple waves of standards, and the delay to the third of these, suggests the 3GPP process is struggling to adapt to the extreme complexity of modern networks.
The first set of Release 15 specs defined the Non-Standalone radio network, a subset of the full standard which was fast-tracked, under pressure from operators led by AT&T, to allow early deployment of 5G networks that could work with the existing 4G core. The decision to allow this option to go ahead was controversial, and critics said it allowed commercial politics to infiltrate the standards process, and resulted in a sub-optimal platform for initial 5G services.
The second set concerned the Standalone network, which requires the 5G core, also included in these standards along with much of the system architecture – but because of the existence of NSA, Standalone commercial roll-outs are not expected at any scale until the early 2020s, by which time Release 16 will have been finalized.
Before that happens, in early 2020, the third set of R15 specs has been made available, filling in some gaps which were left in the preceding bundles. These details are deemed necessary for Release 15, but would have delayed the standard significantly to be included in the initial publication, and were delayed further (from the end of 2018 to April 2019).
The main reason for that delay was, apparently, that engineers were distracted by fixing some bugs in the Standalone 5G NR specs, which were frozen in September. In particular, questions were raised over backwards compatibility of equipment and devices supporting those latest specs, and those based on the earlier Non-Standalone platform.
Now that the late drop has achieved its functional freeze, the new timetable sees the ASN.1 freeze following in June. (The 3GPP defines ASN.1 as covering protocols between the network and the user equipment, or between network nodes, while everything else is considered a ‘functional requirement’.) For Release 16, a functional freeze is now scheduled for March 2020 and an ASN.1 freeze for June 2020. The RAN TSG aims to approve a set of work items for Release 17 in December 2019.
Qualcomm’s head of technology specifications, Lorenzo Casaccia, says that the late drop contains a number of variations on the SA and NSA versions of 5G that operators such as AT&T, and some in Europe, had expressed interest in. In particular, they may provide future stepping stones for transition from NSA to SA for some MNOs, though Casaccia added: “The jury is out on whether anyone will use them ever.”
But while some question whether the late drop was necessary, others argue that, with standards becoming so complex and multi-faceted, the process will never be complete and the 3GPP needs to emulate the IT industry and shift to a more ongoing system of updates rather than a hefty release every year or two. Indeed, a number of companies have already submitted updates and fixes for the Rel15 specs, which 3GPP will need to work on in parallel to its Rel16 projects.
Key elements of Rel16 include many focused on the industrial and IoT applications that are often associated with 5G, but are poorly supported by the broadband-oriented Rel15. These include ultra-low latency and V2X communications. Other important projects include integrated access/backhaul and 5G in unlicensed spectrum.
Some companies are already starting to think about Release 17. Possible work items may include spectrum above 50 GHz, enhancements for drones, 5G Multicast, specifications for industrial sensors, and AI technologies for 5G network operations and automation.
By then, however, will the 3GPP process still be fit for purpose, or will it have to adjust considerably? The number of features that have needed to be standardized for 5G, and the number of submissions for each one, have been greater in 5G than 4G, by 10 times or more, say the companies involved in the 3GPP process. These changes show how complex this new network is, and how there will never be a ‘final’ set of specs to address all the requirements.
This issue is thrown into sharp relief when vendors and operators are also engaging with open source communities like the Linux Foundation’s LF Networking family of projects, which tend to have a faster, more fluid process (though with the downside of suffering from fragmentation).
Peter Clemons, chief designer of critical communications company Quixoticity, and a member of the RAN group, said the delay to the late drop – which will also push out Release 16 by at least three months – stemmed from “sheer workload”, and a lack of coordination between the different 3GPP TSGs (technical specifications groups) working on the 5G standard. As well as RAN, these include the SA (service and system aspects) and CT (core network and terminals) groups.
Clemons told LightReading: “The enormous amount of work undertaken during Release 15 by RAN 1 and RAN 2 [TSG sub-groups] needs to be coordinated with SA 1 and SA 2, with CT also having a really hard job to make sure that rapidly changing standards are backwards compatible and forwards compatible.”