Achieving the reliability needed with such a complex set of technologies may require solutions that go beyond the PHY and MAC layers.
In May of this year, I attended an Emerging Technologies Workshop as part of the IEEE ComSoc International Communications Quality and Reliability Conference. During this workshop, I had the opportunity to present some of National Instruments' work in the 5G space, and to learn how others view 5G’s impact. In particular, I was curious to hear more about reliability as a key performance objective of 5G and the specific use cases targeted by the 3GPP.
While 5G is on a path to deployment—perhaps earlier than many pundits expected—it will not be as pervasive as LTE. In fact, service operators continue to deploy LTE even today. In the 5G context, deployment will take years and perhaps longer than LTE because of the complexity and the flexibility of the standard. Flexibility is what enables the new use cases targeted by the 3GPP for 5G including enhanced Mobile Broadband (eMBB), the Ultra-Reliable Low-Latency (URLLC), and the Massive Machine Type (mMTC) broad categories. Not only must operators deploy the new 5G infrastructure equipment, but the standard as it stands now has so many variations and configurations that the initial deployments will likely be a subset of the immense and comprehensive initial promise.
So where does that leave reliability? Well, everyone who owns a cell phone wants better reliability. Can 5G cure the proverbial, "dropped call?" 3GPP Release 15 includes a few improvements over LTE pertaining to reliability. The flexible frame structure with various options of sub-carrier spacing, modulation, and coding provide a solid foundation at the physical layer. Hybrid automatic repeat request (HARQ) and ARQ are included as well with LTE and will initiate "retries" on corrupted data. In addition, the 3GPP specified new reference signals to improve synchronization to enhance demodulation efficiency, which will significantly improve bit-error rate (BER) and ultimately lead to enhanced coverage.
The HARQ process uses a series of handshakes to retransmit packets. Source: Nokia Siemens Networks
With that said, the 3GPP did introduce some technologies that may prove challenging for enhanced reliability in some use cases. First, the inclusion of mmWave spectrum with phased array antenna technologies will pressure the reliability curve. mmWave signals are highly directional and when blocked, the signal level to the receiver may drop below a signal to noise (SNR) threshold that ultimately "drops" the link. The 3GPP is addressing this issue with beam management and recovery methods but the challenge is not only real but highly probable.
Beam management and recovery schemes must be optimized to prevent "drops." This may not seem all that important for the eMBB use case because our current networks suffer from similar discontinuity in service, which only mildly irritates current users. When mmWave for the eMBB use case works, all users will rejoice. When it doesn't work so well, i.e. "lumpy" data rates, users may be conditioned to this reality from their current experiences, so it may not be a show-stopper.
For the URLLC use case, this type of reliability will surely be unacceptable. As a result, the URLLC use case may only be deployed on frequencies below 6 GHz and use the higher order channel coding. This means the overall data rate may inevitably slide below today’s standard LTE service.
To wrap up, here are some takeaways from my time at the Emerging Technologies Workshop.
- First, operators will need to streamline the collection of network resources to be application specific to ensure that these use cases achieve the results they depend upon. "Network slicing" facilitates this type of resource allocation by implementing user profiles to optimize the 5G network toward specific use cases with a set of performance objectives, and perhaps they may charge accordingly.
Network slicing adapts networks for specific use cases.
- Second, even with the improvements to 5G physical layer, there is no guarantee of better reliability. Network slicing may deliver better reliability for certain use cases but achieving higher levels of reliability greater than that may be difficult to achieve.
- Third, 5G mmWave communication faces several challenges not encumbered by 5G networks operating below 6 GHz.
- And finally, achieving reliability may not be possible without some sort of macro diversity including coordinate multi-point or dual connectivity. There is only so much researchers can do at the PHY and MAC layers, and reliability exploration and evolution may in fact need system or network solutions to achieve the 3GPP goals for reliability.
—James Kimery is the director of marketing for RF, communications, and software defined radio (SDR) initiatives at National Instruments.
- 5G: The road to low latency
- 5G reaches milestone with New Radio specification
- Building 5G networks from the core up
- Millimeter wave beamforming and antenna design
- Spectrum sharing: Another way to better use the airwaves
- 5G Dials in Low Latency, Slicing