Malaysian manufacturing companies realise that significant business value can be derived from shop floor-level data from plants. The problem they face is that current systems and processes in organisations operate in silos and, as a result, there is no seamless transfer of data from SCADA (Supervisory Control and Data Acquisition) systems located on industrial shop floors to ERP (enterprise resource planning) and PLM (product lifecycle management) systems used by most companies to forecast growth and business direction.
To succeed in Industry 4.0 transformation most companies need to be able to break the silos and ensure seamless transfer of data to scale production and profitability by making the best use of data analytics, artificial intelligence (AI) and machine learning (ML).
How companies can navigate this problem and succeed in their Industry 4.0 journey was the subject of a roundtable discussion organised on August 17, 2023, at the Le Meridien, Kuala Lumpur.
More than 30 participants engaged in a lively discussion for more than two hours on the theme: Transforming manufacturing with AWS digital technologies by leveraging significant value from AI, ML and data analytics.
Setting the context of the discussion AWS Enterprise’ Head of Technology, ASEAN Strategic Industries, Dr. Daniel Kearney, noted that at present “we are looking at application stack made up of many monolithic applications” where ERP and PLM systems operate independently with data in silos and produce reports for selected departments and officials.
“And on the production side, on the shop floor, we have SCADA systems, distributed control systems, sensors and actuators all individually operating with their operating system or their applications.
“Communication between information layers is practically non-existent. And we have very disparate protocols that have evolved independently,” Dr. Kearney said.
IT dissolution
He added “At the top of the stack, AWS is driving an IT dissolution... so that the entire process becomes a lot more micro-service-based.
“ERP systems and PLMs are beginning to be broken down and coupled with micro-services getting a lot more freedom of operation,” Dr. Kearney said.
The advantage of doing this is that companies can move from “data access to data-informed, we've got more power users that can now search across different data sites”.
But, Dr. Kearney added, there is still “an isolation” at the bottom of the stack on the OT (Operational Technology) side where we're not getting that level of integration.
“So what the true evolution of tomorrow is, is that we move to a situation where we have an IT-OT border that has been removed or dissolved. This will provide more fluidity to how business people talk to operating managers, how supply chain communicates with manufacturing, how inventory can be managed based on demand signals from customer to the shop floor.”
Dr. Kearney noted that this was exactly what Amazon.com was doing with its systems and processes.
“At Amazon, our North Star is ensuring we always fulfil our customer promise while relentlessly inventing around the customer experience. Amazon realised early on that leveraging data and automation were key to helping us scale our next-day delivery experiences across millions of items globally.
“Today Amazon’s connected manufacturing and supply chain leverages a vast corpus of operational data, including third parties, and combines that data with AWS’ extensive machine learning capabilities to fulfil 99 percent of all customer orders automatically, without human intervention. Those same AWS capabilities are also available to customers to help them ingest, manage and automate operations at an unprecedented scale,” he said.
Dr. Kearney noted that one of the things about Amazon is that "when we talk about end-to-end visibility, and we talk about data, if you can't measure it, you can't manage it.
“It's all well and good to have the capability to aggregate everything. But your teams are also operating across the silos and your teams are also now trying to communicate and understand all of the highly interdependent touchpoints across such an end-to-end visibility point… Now, the speed of business depends on people at the very edge, whether it is on the production line or in the field in a delivery van, having access to that data to make a real-time fast business decision, because that's the minimum bar that customers expect today.
“So that's what differentiates businesses in the real world and what sets you apart from other customers,” Dr. Kearney said.
Dr. Kearney said it was important to understand what is the question or problem you are trying to solve in Industry 4.0.
When you have that rationale in your thought process, you (have to) start to think about, well, if I'm trying to expedite my business value, what does that look like? That in a way is the North Star that should guide companies on their Industry 4.0 journey, he said.
AWS partner o9 Solutions’ Vice President, Industry Solutions, Schalk de Klerk, who was one of the speakers at the event, noted that manufacturing companies need to be “very clear not only on the use cases they want to solve now but also the pain points and challenges they've got today and that they know, two years, three years, four years down the track how they want to move up that maturity curve. So I think what is the key is to understand what good looks like in your context in your environment, with the potential limitations that you have.
“But I think that's only one part of the equation. The other part is to pick an area where there's the greatest value leakage today, or to put it in another way, where the greatest challenge exists, whether it's on forecasting accuracy, which leads to higher inventory levels, or if you've got an internal bottleneck in your plans… Start there, focus there, but start small, and start with the simplest stuff and add the more advanced type capabilities over time,” de Klerk said.
Find the right partner
He added that it was important to find the right partner to do this. “Obviously with AWS and its partners, you're able to do that."
de Klerk added: “The last point that I would like to add is that there is a sort of concern or hesitation that we get pretty much anywhere we go in any country, any sector and that is about data. Companies typically want to go down a data improvement type exercise, before they look at deploying enterprise-wide platforms.
“I think realistically, from our perspective, you will never have perfect master data, you will never have perfect planning data. The simple reality is today your planning data may contain Excel (data) with different people in different departments... It's pretty easy to standardise that and put that into a common platform.
“And then to make that trade-off call between what needs to be in your ERP system on your planning platform as well. So I think, just keep in mind, if data improvement needs to occur, use the use-case that you've chosen, whether it's the largest value to an essential process that needs to be unlocked, as the catalyst to go and fix that data and the processes behind it,” de Klerk said.
“Don't go from an end-to-end data improvement perspective, really use that one use case to help drive what datasets you need to improve as well to sort of focus capabilities as well,” he added.
de Klerk said it was “really about having the platform, the data, the people all aligned, but then making those decisions with the best analytics and AI, ML type capability that you've got available to you to make those decisions far faster and more frequent”.
Transparency is key
Pinspace, a warehouse robotics start-up, CEO, Chua Di Ken, noted that from a warehousing perspective inventory is your money. “So having that level of transparency to plan your inventory better, knowing when you need to purchase, how much you should purchase, by having a fully transparent warehouse… that can change your mode of operation,” Chua said.
Giving an example, he said, for example, if a company does not have full transparency in its warehouse, it would have to, maybe, buy inventory every month as that is the only term its systems can cope with.
“But let's say if you improve your operation efficiency, have more transparency and less leakage, then the mode of buying inventory will change, and that improves the efficiency not only in your warehouse but also in the procurement side.
“So having transparency at onsite will not only affect the onsite in terms of ROI (return on investment), it will also affect the front end and the back end later on as well,” Chua said.
Talking about Pingspace’s business model he said that its two main components were robotics, which is building robots for the warehouse, and the other was fulfilment at its warehouse sites.
The company does a lot of experimentation to build better robots and collect data that helps in the fulfilment side, he said.
“But progressively for us, it is not about the robotics itself that provides more value. It’s the data we use to further optimise our algorithms… data is our North Star which we try to improve,” he said.
Picking up this point Dr. Kearney added that for AWS data was like a flywheel. The more data we inject in the flywheel gets more momentum and becomes more valuable and hence becomes future proof, he said adding: "That's something we have at our (AWS) core… it's a very important part of our business".
AWS offerings
Based on interactions with the audience which was keen to know more about the services that AWS could offer, particularly to small and medium-sized enterprises (SMEs) Dr. Kearney shared the various key offerings that AWS had for customers:
AWS for Industrial is an intentional focus on meeting customers where they are with fit-for-purpose industrial solutions. AWS for Manufacturing is one of 19 industry verticals where AWS has tailored solutions and frameworks to help customers get started quickly on AWS and begin capturing immediate benefits. Customers can get started on AWS quickly in 3 ways:
- AWS Industry Solutions like the AWS Industry Data Fabric help customers create the data management architecture that enables scalable, unified, and integrated mechanisms to harness data as an asset.
- AWS Marketplace: is a curated digital catalogue that makes it easy to find, test, buy, and deploy the third-party software customers want, with the simplified procurement and controls you need.
AWS Partner Network (APN) is a global community of over 100,000 partners that leverages programs, expertise, and resources to build, market, and sell customer offerings.
AWS open architecture allows customers to tailor their IT and OT environment with open API architectures that don't require any punitive licencing agreements. This approach also the customer to stand up in the right environment that can evolve with their business needs with the best cost-performance and remain compatible with the ecosystem of third-party vendors and providers.
