Expectations for 5G edge cloud infrastructure in industrial and automotive - Embedded.com

Expectations for 5G edge cloud infrastructure in industrial and automotive

The edge cloud is integral to virtual 5G infrastructure design and to operators’ ability to offer high performance services to advanced enterprise customers. In a recent survey conducted on behalf of Wind River by Heavy Reading, 38 percent of operator respondents identified the edge cloud as “critical” to their 5G strategy and 47 percent as “important.”

Enterprise and Industrial 5G Edge Cloud

The 5G system is designed, in part, for enterprise focused use cases. The intent is to apply 5G technical capabilities across diverse industries, which in turn drives new revenue opportunities for service providers. This applies, in particular, where there is a need for ultra-reliable low latency communications (URLLC), which cannot be supported over LTE.

Examples are cyber physical systems with deterministic performance requirements, such as industrial robotics or remote control of machinery where larger round trip latency cannot meet real time system requirements. There are also other demanding applications that cannot be served with existing wireless technologies as they are challenged by scale as well as latency. Examples include autonomous vehicles (ground and airborne), machine vision, virtual and augmented reality for industrial applications, telemedicine, and so on.

According to a recent Heavy Reading study on enterprise and industrial users’ perspectives on 5G, just over a quarter of respondents appear sold on 5G already, saying “performance advantages will force its use” where wireless is needed in industrial networks.  The majority, however, say that the claimed, theoretical advantages of 5G still “need to be proven in practice.” They are favorable toward the technology but are not yet convinced. This accords with the proof-of-concept (PoC) activity and trials large industrial companies have been running as they seek to integrate 5G into their production environments, as well as typical early adopter paradigms.


Figure. In a recent study, participants responded (n=43) with their view view on 5G for industrial use-cases, relative to LTE and WiFi? (Source: Heavy Reading)

Additionally, analytics applied to monitoring of edge cloud infrastructure is critical to day-to-day operations. When asked about the importance of various types of embedded edge cloud monitoring, respondents believe embedded monitoring is more important to infrastructure than it is to the applications themselves or to deriving business insights.

Automotive Sector and 5G

The automotive sector is going through a period of rapid and deep change, with the connected vehicle trend (telematics, navigation, infotainment, etc.) occurring at the same time as advances in driver assistance/automation and mass-scale electrification. Mobile networks, in the broadest sense, are important enablers for each of these. For example, the 3GPP effort has a dedicated 5G development track for automotive. There is also activity in the IEEE on the Dedicated Short-Range Communication (DSRC) standards group, as well as in the satellite sector.

Emerging as popular solutions, the 3GPP C-V2X initiatives started in LTE and is now gaining pace in 5G. A key part of the C-V2X concept is the road side unit (RSU), which is a highly distributed edge cloud node from a hardware and cloud infrastructure software perspective. In time, as deployment gains pace, RSUs are likely to see significantly denser deployments than current RANs, making it an important market for “far edge” infrastructure platforms.

In terms of which C-V2X applications will lead deployment, vehicle-to-network is the clear leader, according to the Heavy Reading survey, with 56 percent of the response. This refers to vehicles communicating with cloud services for information like real-time traffic routing and is the simplest V2X service.

One of the biggest sources of hyperbole in the early years of 5G development was equating autonomous vehicles with 5G. The intent was to show how 5G could help advance connected cars and autonomous driving. However, these statements were often extrapolated out of context and taken to mean that autonomous driving could not happen without 5G. This is incorrect because autonomous is developing without 5G and because it is, and always was, unrealistic to assume that 5G coverage would exist everywhere an autonomous vehicle might travel. According to the Heavy Reading study, the largest group of respondents (36 percent) believe that 5G is a “requirement” for L4 autonomous driving.

Mobile network coverage is generally good, but it is far from ubiquitous. In geographically large markets, in particular, it is hard to cover everywhere with terrestrial systems. There are now several non-terrestrial systems in development for wide-area broadband access, with low orbit satellite systems showing promise. At face value, a two-way satellite could be a good solution for vehicle connectivity as an adjunct to cellular and for use cases where low latency is not vital.

When asked if they had plans to integrate satellite connectivity into their 5G automotive offering, respondents’ answers were split almost equally with 34 percent answering “yes,” 35 percent answering “no plans” and 31 percent saying they “don’t know.” The “no” camp might be explained by factors like the additional cost and business relationships satellite would require, by the fact that satellite for two-way mobile vehicle communications is less mature than 5G for the mass market, and by the lack of support for low latency services. The “don’t knows” speak for themselves.

Conclusion

As LTE or Wi-Fi cannot meet the performance requirements of industrial use cases, the perception is that 5G will be superior for industrial applications, thus enabling many of them. Clearly, the emergence of these new industrial and enterprise edge applications is also forging the way for the 5G edge to be fully virtualized, as an enabler to host these applications. We are seeing a transition in contemporary expectations that assume the emergence of autonomous vehicles, telemedicine, Industry 4.0 automated factory and industrial internet of things (IIoT). It will be interesting to watch these use cases converge with the ramp up of available 5G infrastructure.


Paul Miller is Chief Technology Officer at Wind River, responsible for the company’s technology strategy. With nearly three decades of telecommunications and advanced technology leadership at both large companies and successful startups, he is currently focused on Wind River’s edge virtualization and AI solutions, including Wind River’s market-leading 5G Cloud offering based on StarlingX.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.