ACHTUNG! THIS IS THE LAST CHANCE TO JOIN THE WEBINAR ON CRITICAL 5G EDGE WORKLOADS ON SEPTEMBER 27 (TOMORROW). JOIN HERE – IT’S FREE, AND NOT-TO-BE-MISSED!
You know the drill by now; you know the questions. Here is the response from AWS, with the latest postcard from the edge. Unsurprisingly, the message from Amazon Web Services (AWS), about how to divvy-up critical Industry 4.0 workloads between the edge and the cloud, puts focus on consistency across the whole compute continuum. The firm has it covered, it reckons, and it is hard to disagree – if your enterprise is an AWS cloud shop, already.
Because the news out last week saw AWS introduce a new managed edge service under the banner AWS Dedicated Local Zones, which offers to take responsibility, effectively, for the whole Industry 4.0 edge-setup (and provisionally go up against system integrators and mobile ioperators in the edge management game). The new service means enterprises can build their own compute infrastructure on their own privately-owned premises for their own exclusive usages, likely networked on new private 5G gear, and hand management duties over to AWS.
The service is geared for public sector and industrial customers running critical workloads in private edge setups, including those jigsawed into new private 5G network deployments in order to stand-up new Industry 4.0 applications. In conversation with RCR Wireless, Jan Hofmeyr, vice president for the firm’s ‘elastic compute cloud’ (EC2) proposition at the industrial edge, untangles the whole critical 5G edge in the context of his company’s own strategy.
Over the past couple of years, AWS has moved its cloud functions into the wild and wooly ‘edge’ estate, running all the way from regional edge data centres (AWS Regions), to the MEC-edge in mobile networks (AWS Wavelength), to the urban edge in business parks and cities (AWS Local Zones), right down to the enterprise-edge in factories and buildings (AWS Outposts), and even to the live and operational edge in devices and gateways (AWS Snow Family).
Since the interview, these various service solutions, which each embed the company’s cloud functions across different stages of the edge continuum, are now complemented by the company’s new enterprise edge management service, which offers to orchestrate its own storage and compute services, plus its partners various Industry 4.0 analytics and enablement applications, in enterprise-owned (non-AWS) data centres, wherever they are sited.
Hofmeyr references a whole bunch of other AWS bits-and-pieces, available to plug into its edge package: its data visibility (SiteWise Edge) and content delivery (Amazon CloudFront) systems; its data processing and insights tools (Amazon Redshift, Amazon EMR, AWS Lambda); its double-punch private 5G offers (AWS Private 5G, Integrated Private Wireless), and other networking services (AWS Direct Connect, AWS VPN, AWS CloudWAN).
At one point, he breaks-out the Snow Family services, which afford a way for critical 5G edge systems to operate in standalone fashion, to meet the demands for “environments like oil fields, mines, and shipping carriers” to operate in venues without a backhaul to the internet, in cases where this (or might be risk of) ‘denied, disrupted, intermittent, or limited’ (DDIL) connectivity. These are EC2, EBS, S3, IoT Greengrass, Sagemaker Edge – ad infinitum.
Indeed, there are too many products to make easy sense; all of these references have been stripped from the below transcript to keep a clearer focus on the broader challenge at the 5G edge – and just to ease the bombast (some). And Hofmeyr’s responses are excellent, making the case for consistency right across the edge-cloud continuum, and arguing – as AWS always does (very well) – for problem-first/technology-last consultancy and collaboration.
It is not the last word on the critical 5G edge, but it makes a good read. Hofmeyr adds, as a final thought, that data governance, as dictated by regional governments and industry regulators, must inform every architectural decision. “It is crucial to discuss data governance and security… unique to each industry sector. Customers have to weigh their business challenges against the benefits technology provides, from the cloud to the edge – to choose the right solution.”
All the answers below are from Hofmeyr.
…
How should one rationalise the industrial usage of edge and cloud (and in-between) compute resources for processing critical IoT / OT workloads? Is it possible to answer this, firstly, in terms of horizontal applications, and secondly in terms of vertical industry sectors?
“Customers are looking for solutions that provide the same experience from the cloud to on-premises and edge applications. Regardless of where their applications may need to reside, [they] want to use the same infrastructure, services, APIs, and tools. [They] want a continuum of consistent cloud services that extends the cloud into customer premises [and] remote locations – allowing them to operate seamlessly across many environments.
“Companies do not want their teams to be distracted by learning new tools or dealing with new operations, and they want builders to be free to focus on building, [in order to] accelerate innovation, increase cost and operational efficiencies, and adopt cloud security wherever their applications and data are. [After all] the goal is to make data-driven decisions that lead to improvements in efficiency and productivity.
“Industrial customers want enterprise-wide visibility, which requires collecting data at the edge and processing data at the edge – to facilitate real-time monitoring and actions on issues that need to be addressed on the shop floor. But they are also collecting large amounts of data at the edge which can be brought to the cloud, and contextualized and standardized to provide access to a single view across production lines and plants globally.
“AWS offers scalable and flexible compute, storage, and networking services that are very critical and useful across different industries. These are suitable for industrial applications that require extensive computational power, data storage, and complex analytics. [We] provide compute, storage, and data processing capabilities at or near the edge of the network to reduce latency and improve real-time processing for IoT devices.”
“[But ] customers must consider their unique requirements… Where low latency is non-negotiable, such as real-time data processing and analysis, [edge] solutions can minimize round-trip delays for data transfers and responses. AWS offers scalable compute and storage… for applications with unpredictable data volumes or processing requirements.”
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 – and, indeed, when in-between MEC-style regional data centres might be used? Is it possible to answer this also in terms of horizontal applications, and in terms of vertical industry sectors?
“While there isn’t a one-size-fits-all rule-of-thumb for deciding where to retain data across the edge and the cloud, there are several factors to consider across horizontal and vertical applications. The optimal location for data retention often depends on the specific use cases, data characteristics, performance and scalability needs, regulatory and compliance requirements, and also cost considerations.
“Where a [regional cloud] is not close enough to meet latency or data residency requirements, customers need [compute] infrastructure closer to their data source or end users. They can run workloads with low latency requirements [at the edge and] seamlessly connect to the rest of their workloads running in [a regional cloud]. With massive data volumes… it may be practical to preprocess data locally and only send relevant data to the cloud.
“AWS offers limitless storage resources for applications with unpredictable data volumes to accommodate spikes in [traffic and] computational demand… There are several use cases [that] leverage the [whole] edge-to-cloud continuum… They span multiple industries. By working backwards from the customer’s requirements, we can determine the best fit for the workloads.”
Are there Industry 4.0 scenarios where the data will stay on the edge, all the time, separately of a central cloud? Can you give examples?
“There are Industry 4.0 scenarios where the data will [always] stay on the edge. In manufacturing,IoT sensors that reside inside manufacturing equipment continuously collect data on environmental attributes such as temperature, pressure, and vibration. This data is processed at the edge in real-time to make quick decisions to ensure manufacturing processes remain within optimal range and to minimize any possible downtime.
“Environmental data is captured at the edge continuously… [But] industrial customers [only] care when it goes above or below a certain threshold – so they can act on it. [They do not care about] all of the data points captured in between [the threshold measures]… Smart farming is another example where computer vision is deployed at the edge… to automate the labour-intensive practice of counting cattle, saving farmers time and money.”
Do you have a matrix / can you advise on a matrix to show which use cases / applications (in which industries) will be handled at the edge and in the cloud, and at points between?
“There are several use cases best-suited to leverage the AWS edge-to-cloud continuum: delivering low-latency applications, migrating and modernizing on-premises and edge applications, meeting data residency requirements, and processing large volumes of data locally. These categories are evolving based on our customers’ requirements.”
What is the tradeoff between low-latency private-5G and low-latency edge compute? Do you have a guide for how these two technologies combine to max-out network-and-computing performance? Can you say (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?
“Both solutions approach the same goal from different angles. Private 5G networks enable local data communications between devices, and handle reliable and rapid communication with a high density of connected IoT devices. They can accommodate a large number of industrial devices without impacting latency, provide coverage over large areas, and offer inherent data security and isolation due to their dedicated nature – thus minimizing unauthorized access.
“Low-latency edge computing offers data processing near the data source, bringing it closer to the end-users and enabling real-time data processing and more control over how data is processed. In essence, both technologies are mostly complementary, so trade-off between private 5G and low-latency edge computing involves optimizing for connectivity, processing speed, scalability, and resource constraints – depending on the use case.”
All of which, 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?
“We work closely with partner companies to solve all of the different Industry 4.0 requirements from customers – including independent software vendors, systems integrators, and communication service providers. Our new Integrated Private Wireless [offer] emphasizes the importance of collaboration with partners… Industry 4.0 deployments span a diverse set of technologies… [and], more often than not… require multiple vendors and seamless integration… Which underscores the importance of ecosystem consultancy and collaboration.”
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
Postcards from the edge | Edison on the see-saw gains between 5G edge and cloud
Postcards from the edge | Cloud-to-edge consistency is everything, says AWS
