The 5G New Radio: 3GPP standards progress

The 5G New Radio is certainly new. It is perhaps not new in the same obvious way as TDMA, CDMA and OFDMA were in previous generations. However, the pervasive changes required to support all the new 5G use cases more than justify this exalted status.

Only a few years ago, our industry was largely skeptical about the value of anything new coming along in a next-generation radio. These doubts have gradually diminished as new approaches and coding schemes have been proven relevant to emerging use cases.

An important milestone was reached back in March when 3GPP published its first study item reports on New Radio (NR). In these reports and the standardization work that has happened since, it is becoming clear just what this new radio is going to look like.

However, one question remains, will this new radio really be a true paradigm shift away from 4G or simply more of an overhaul and a refresh for a new generation.

A brief reflection on the anatomy of a 4G radio

4G, or LTE, was conceived and designed for the optimal support of the mobile broadband use case, i.e. video support. The radio or radio stack is composed of multiple protocol sublayers, namely, PDCP, RLC, MAC and PHY. The highest layer PDCP is responsible for processing IP packets and provides services like compression, ciphering and integrity protection. RLC is the next layer down and handles functions like segmentation/concatenation and error controls through procedures like Automatic Repeat Request (ARQ). The MAC layer is responsible for scheduling, multiplexing, a Hybrid ARQ process and all management of transport blocks transferred to and from the PHY layer.

The PHY layer is the layer that is most traditionally associated with the radio chain and includes the core radio functions: channel coding, modulation, MIMO and multiple access procedures. The channel coding scheme selected for LTE was Turbo codes. QPSK and up to 256QAM modulation (depending on cell topology) is supported. SC-FDMA and OFDMA multiple access techniques are supported on the uplink and downlink respectively.

One important thing to note here is that the LTE multiple access schema was designed with a so called fixed numerology. This refers to the fact that LTE defines only one sub carrier spacing of 15KHz. This is basically the lowest form of radio resource definition in the frequency domain in LTE. This fixed approach chosen for LTE was perfectly reasonable if we reflect on the single use case that dominated the design. However, by now you should know that 5G is targeting a much broader set of use cases, and flexibility (not fixed or rigid) is perhaps its most regularly used key word.

So, what is new about the 5G New Radio?

The first and most obvious difference is that there will be distinct frequency bands ranging from sub 6GHz all the way up to the much higher millimeter wave (mmW) bands. What is not so different is the multiple access and modulation. Like 4G, OFDMA has been selected with the nuance that it may be applied this time on both uplink and downlink. The option for the UE to still transmit on the uplink using a more efficient single carrier method (DFT-S-OFDM) is preserved.

The key new element here is that a flexible numerology has been introduced into the design. The 5G NR continues to support a basic sub carrier spacing of 15KHz but will now allow this to be expanded in multiples of 2. The main motivation for this is latency reduction. Increasing sub carrier spacing enables shortening of the OFDM symbol duration, which in turn enables lower latency communication. Further, transmission with different numerologies can be multiplexed in both TDM and FDM modes enabling highly flexible resource sharing arrangements across different traffic classes/use cases.

A new mini-slot, greater emphasis on spatial methods & new channel coding

5G NR introduces a new “mini-slot”. In LTE, a slot is defined as seven OFDM symbols. The new mini-slot being introduced in 5G NR may occupy as few as two such symbols. The devil is in the details of course but this mini-slot design is really all about enabling so called ultra-reliable low-latency communication (URLLC) use cases where it is in some cases very desirable to switch between transmit/receive in the uplink/downlink more quickly. The motivation here is again all about reducing latency.

MIMO has played a key role in 4G to boost the peak data rate, and this trend is set to continue in 5G especially in the mmW bands where transmission and reception using narrow directional beams will be essential. Compared to 4G, it is fair to say that there is a much greater emphasis on the standardization of procedures in the spatial domain. This will introduce an extra degree of freedom as radio resources in this domain will be much more readily exploitable for new performance benefits.

Turbo codes have been a cornerstone of both 3G and 4G and for a while, it was assumed that nothing could replace them. Not so in 5G NR. LDPC codes have been selected for the enhanced mobile broadband (eMBB) use case based on their proven superior performance for larger block sizes, particularly relevant to this use case. In a surprise turn of events, Polar Codes have been selected for the eMBB control channel due to their perceived superior performance over Turbo codes for small block sizes. The selection of these new codes, while only for one of the 5G use cases so far, represents a major change from 4G.

And the higher layers?

As you might expect with so many changes in the PHY, there are many necessary responses in the higher layers to support them. Many new features are also being introduced to provide further support of the pervasive latency reduction theme in 5G. For example, in the RLC layer, to reduce processing time it will now be possible to disable concatenation. In the MAC, new enhancements are being introduced to the HARQ procedure to speed up retransmissions. A duplication feature is being added to PDCP to increase reliability. This will allow the same PDCP Data Units to be sent over different carriers when carrier aggregation is in operation reducing the number retransmissions required.

A new user plane layer is being added above PDCP called the Service Data Adaptation Protocol (SDAP). This is a new dedicated layer in the New Radio that is being added to support the much more complex Quality of Service functionality in 5G.

So, is the 5G New Radio really that new?

The 5G New Radio is certainly new. It is perhaps not new in the same obvious way as TDMA, CDMA and OFDMA were in previous generations. The selection of OFDMA in 5G is perfectly reasonable given how successful a solution it has proven to be in supporting mobile broadband in 4G. We should never forget that enhanced mobile broadband will very likely remain the most important use case in 5G. However, the use cases in 5G are much broader than this and the pervasive changes being introduced across the stack to support them all more than justify this New Radio classification. What is not clear yet is just how many of these changes will in time make their way back into the ongoing evolution of LTE. This would be perfectly normal and would certainly not be the first time that the first customer for new generation innovation is the old generation.

Copyright © 2017 IDG Communications, Inc.

7 inconvenient truths about the hybrid work trend
Shop Tech Products at Amazon