As part of the ongoing Postcards from the Edge series, RCR Wireless caught up with Dan Hays, principal at PwC’s strategy consultancy division Strategy&, to discuss how the Industry 4.0 ecosystem might impose some kind of industrial order on the new digital chaos at the critical edge, especially as novel private cellular systems start to be (slowly) levered in. How should complex industrial workloads be divided between edge and cloud systems?
What stays and what goes in terms of industrial data? How does latency go to single-digit milliseconds as computing and networking come closer to the edge – and where should enterprises, and their suppliers, halt the sliding scale to hit a sweet spot between system performance and infrastructure costs? RCR had the questions, Hays had the answers. The full interview is printed below; all the responses are from Hays.
…
How should one rationalise usage of edge and cloud (and in-between) compute resources for processing critical Industry 4.0 workloads?
“The concept of edge computing is not really a particularly new one, though in the age of cloud and private networks, it has taken on new meaning. In modern times edge computing is really being conceived of as a means to address the cost and performance limitations of communications networks, particularly those that connect remote locations to cloud infrastructure. Establishing edge computing capabilities can increase the responsiveness of decision-making systems to the data that is feeding them and lower the cost of sending data and decisions back and forth between the cloud and the locations that it serves.”
Is it possible to rationalise edge usage, firstly, in terms of horizontal applications, and secondly in terms of vertical industry sectors?
“This is especially true for applications that require real-time data, including manufacturing and operational control systems, security systems, traffic/congestion management, and even healthcare monitoring systems. These can apply to virtually all industries, though manufacturing, transportation, and healthcare are particularly notable.”
Is there a rule-of-thumb for when IoT / OT data should be retained at the on-site edge, and when it can be diverted by the cloud?
“While rules of thumb are still emerging, drivers of edge computing tend to focus on use cases where the desired latency of control systems needs to be less than the round-trip time to transmit data to and from the cloud. And, furthermore, the value of the controls needs to be greater than the incremental cost of establishing the edge computing capabilities.”
When should in-between MEC-style private regional data centres be used? And can this also be answered in terms of horizontal applications and vertical industry sectors?
“Private regional data centres are a good option for many use cases and environments, especially where data volumes are modest and cost sensitivity is high. Regional data centre models can be expected to apply in environments where infrastructure is particularly distributed such as oil and gas wells, electric infrastructure, and vehicle traffic management. It will also be most effective where less real-time applications are planned, such as data logging or slower response-rate control systems.”
Are there Industry 4.0 scenarios where the data will stay on the edge, all the time, separate from the public internet? Can you give examples?
“In reality, much of Industry 4.0 has little to no need to ever touch the public internet. This is especially true of many IoT applications which are being used for monitoring and control of remote systems. Much of this traffic can be effectively contained within private local and/or wide-area networks.”
Will all-edge private 5G networks, where cellular is required, always be used in these cases?
“Private 5G is only one of the options for edge-centric connectivity. Given the lower cost and greater availability of devices, private 4G continues to be a very strong option for many networks. And even older, lower-bandwidth LPWAN technologies such as LoRaWAN are continuing to grow in adoption for less traffic-intensive applications and services.”
What interfaces/access will these edge networks have with the public internet and public cloud? When, why, how?
“Even in situations where edge networks are prevalent, the resulting data is likely to be warehoused and further processed in more conventional public cloud environments. This will require private networks to be interfaced with public ones in order to take advantage of the associated economies.”
What is the tradeoff between low-latency private-5G and low-latency edge compute?
“Low latency private 5G networks are expected to work hand-in-hand with low latency edge computing. Whereas the network is typically all about moving data back and forth from sensing and control devices, the edge computing infrastructure is focused on rapidly performing the required analysis and decision making in the midst of the data’s round-trip. In reality, the edge computing infrastructure is competing more with backhaul communications to see which can deliver the responsiveness and cost required for the given application.”
Do you have a guide for how these two technologies combine to max out network and computing performance?
“It is really a question of how the three parts add up – the connection from the endpoint to the computing, the processing time in the computing infrastructure, and the return trip from the computing resources back to the endpoint. Looking at the time and cost associated with each of the three pieces allows decisions to be made about which architecture to be used for each.”
Can you say (again, rule-of-thumb / on paper) how latency reduces with 5G from the cloud to the edge, in terms of networking and computing, when employed separately and together?
“There is really no rule of thumb, unfortunately. It is a complex set of calculations based on the architecture of each solution.”
All of this, presumably, makes clear how complex Industry 4.0 deployments can be, and how important ecosystem consultancy and collaboration is. Is this right? Can you comment?
“Industry 4.0 is really about automating and connecting the world around us. With millions of combinations possible based on use case, use environment, applications, and infrastructure, there is no single recipe for success.”
For more on this topic, tune in to the upcoming webinar on Critical 5G Edge Workloads on September 27 — with ABI Research, Kyndryl, Southern California Edison, and Volt Active Data.
All entries in the Postcards from the Edge series are available below.
Postcards from the edge | Compute is critical, 5G is useful (sometimes) – says NTT
Postcards from the edge | Cloud is (quite) secure, edge is not (always) – says Factry
Postcards from the edge | Rules-of-thumb for critical Industry 4.0 workloads – by Kyndryl
Postcards from the edge | No single recipe for Industry 4.0 success – says PwC
Postcards from the edge | Ultra (‘six nines’) reliability – and why it’s madness (Reader Forum)
Postcards from the edge | Private 5G is reshaping the Industry 4.0 edge, says Nokia
