Cloud solutions enable manufacturers to seize benefits around cost reduction, flexibility, and new digital and smart capabilities. Companies seeking to make the most of all of these will likely have to review and rethink their manufacturing operations management strategies. Find out how to equip executives with key pointers for how to best start this exercise and ensure cloud success.
<<< Start >>>
Maximize your business efficiency with cloud in manufacturing.
<<< End >>>
Historically, MOM applications have been deployed either as instances inside manufacturing plants or in centralized on-premise data centers. The decision for local (inside manufacturing plants) deployment has been mainly driven by security and availability concerns associated with external network links.
However, this reality has changed. Robust network links are more affordable nowadays in many locations and public cloud providers have developed hybrid architectures to address those concerns. In addition, user exposure to cloud applications such as ERP, PLM, Office365, ServiceNow, and Salesforce has demonstrated that cloud applications can indeed provide good experiences to end-users.
We spoke to Thiago Martins, a Managing Director in Accenture’s Industry X group and the global leader of the firm's Manufacturing Execution Systems practice, about the state of cloud in manufacturing and what benefits it offers.
Hi … Thanks for talking to us today. To start, can you tell us a little about the state of cloud for manufacturing?
Thiago Martins: Hi … Well the good news is cloud providers have launched many different technology components that can help manufacturers implement use cases that, not so long ago, were deemed too complex or too expensive. Voice recognition, image processing, machine learning, and geoprocessing are just some of the new tools in the manufacturer toolbox that can help operators and supervisors make the right decisions during their shifts. So, the first point to consider is: cloud has been enabling different ways of working in the shop floor.
This is perhaps the main driver for cloud adoption, beyond the well-known IT efficiencies such as IT footprint reduction, better price-to-performance ratio, enhanced security, etc. Cloud is no longer a futuristic thing, especially in regions where the available network infrastructure is resilient enough to guarantee uptime and performance. It is important to consider, however, that these trends are not being seen at the control and supervisory systems levels (ISA S95 Level 2), which remain close to plant equipment in the shop floor.
And what kind of benefits does cloud enable for manufacturers?
Cloud enables two groups of benefits: business efficiencies, and IT efficiencies.
From the Business side, for example, cloud allows for better collaboration by eliminating data silos and by enabling subject matter experts to collaborate across plants in different geographies. Cloud providers represent the main access gate to software innovation such as Artificial Intelligence, Analytics, Industrial Internet of Things, No-Code/Low-Code platforms, Mobility, Geo processing, and many others. By using these services, manufacturers can, for example, use crowd-sourced machine learning algorithms to help explain and predict equipment failure. They can use advanced analytics to share interactive dashboards with other companies across their supply chain, etc.
<<< Start >>>
Business efficiencies and IT efficiencies benefit from cloud.
<<< End >>>
Going to the IT side, cloud hosting means a reduced IT infrastructure footprint at manufacturing sites, the ability to scale capacity as needed (instead of oversizing on-prem infra), better disaster-recovery capabilities, and automated administration, to name a few.
However, when it comes to manufacturing, there are risks to be managed, and here is where some of the initiatives to implement cloud in manufacturing encounter some roadblocks.
And what are the risks?
What is different when it comes to the shop floor is that unlike many other corporate applications, manufacturing systems are usually interconnected with plant equipment. This can pose additional challenges to remote hosting due to the introduced failure points represented by external network circuits, as well as their latency. Many of these applications are mission critical to manufacturing, which means that the cost of unplanned outages can be high. In addition to this, security and data privacy are big concerns as manufacturing data usually reflects company’s intellectual property.
These risks are usually well known by the organizations. However, ways of avoiding/mitigating them are not so well known because most of the remedies are new. Developments made in several aspects of the technology stack have turned cloud into one of the safest and most reliable hosting options. Cloud providers offer multiple services to protect the application and data they store.
Examples include directory services like Azure AD for authentication and authorization support, application-defined firewalls to enforce connectivity policies, and distributed denial of service prevention such as AWS Shield. Not to mention hybrid components such as AWS Outposts, Google Anthos, and Microsoft Azure Stack that allow cloud services, APIs, and tools to become available inside of the manufacturer’s very own data centers.
<<< Start >>>
Risks are usually well known by the organizations, ways to avoiding/mitigating them are not so well known.
<<< End >>>
So how are manufacturers approaching cloud then?
I am glad you asked that because there is a big misconception when this discussion comes up. I am not seeing, and do not expect to see in the short term, manufacturers move their control systems to the cloud. The cloud adoption typically happens first at the ISA95 Level 4 (corporate systems), and then it reaches some Level 3 applications. And at level 3, we see two things happen: manufacturers are bringing new capabilities powered by cloud to support their operations; and manufacturers are migrating existing assets to cloud.
It sounds like you are making a distinction between cloud migration and cloud adoption?
Exactly. You don’t need to migrate your systems to the cloud to reduce cost, increase safety, maximize yield, etc. Cloud is a significant accelerator to address gaps in the services that the IT or engineering organizations currently provide to manufacturing. Benefit can be achieved by introducing new capability without having to migrate existing assets.
Of course, cloud adoption, be it done via introduction of new services or migration of workloads to external providers, require strategy. Because cloud providers make experimentation easy, we see a proliferation of well-intended and unfinished small “pet projects.” These are essential to innovation, but without a consistent strategy, most of these initiatives end up not scaling, or not delivering the expected outcomes.
<<< Start >>>
<<< End >>>
And what should manufacturers do to avoid these traps?
The journey should start by identifying business outcomes that you want to achieve. Once the objective is clear, the next step is to evaluate existing applications. Which ones should be retired? Think about those reaching end-of-life or no longer business-relevant. Which should remain as is? Include here the ones that are used by a very small subset of users, that are too complex to migrate, or whose migration does not offer a compelling business case. Which ones should be replaced? Think about those that can be substituted by commercial off the shelf packages or those with high functional overlap with other existing applications. Which ones need to be changed, be it a small patching or a complete refactoring? Which ones can be wrapped into containers or VMs and lifted-and-shifted to cloud? And finally, which ones can become obsolete if the plant operates in a completely different way, if the work processes is re-imagined?
As manufacturers go through this exercise, we recommend that they optimize for drop-and-shop vs. lift-and-shift. By doing this, manufacturers can leap-frog generations of applications and move to fit-for-purpose and future-proofed solutions.
So, this requires a great deal of technical knowledge, right?
Correct! For example, let’s talk about sourcing model. Many companies using cloud today have a multi-cloud strategy with 5+ cloud providers. This means that as you evaluate what to do with your workloads, applications might land at AWS, GCP, Azure, Oracle, or other platforms as virtual machines, containers, serverless, or even be replaced by other options.
Of course, you want to make decisions that can be revisited in the future, and here is where the portability concept comes in handy. Unlike vendor lock-in scenarios when software was tied to proprietary hardware and operating systems (e.g. VAX/VMS), exiting a cloud platform today is less cost prohibitive due to the increased portability of cloud applications (think about containers).
As you said, understanding of the ever-changing technicalities is important, but more than that, understanding of business needs and manufacturing process is a must to make the right decisions.
<<< Start >>>
<<< End >>>
And it is of course a decision that goes beyond the technical aspects, right? It seems that this is getting into the realm of strategy and organization.
This is all about the strategy. What is the business outcome that you want to achieve? Is cloud the right solution? If so, how to adopt it? But then, when we look at this holistically, you need to understand that cloud adoption is not a project. There is no start and end to it. Once you are in the cloud, there are things that the organization needs to do to sustain the benefits.
Let’s talk about the operating model. The traditional model of managing capacity by purchasing and running hardware doesn’t work in the cloud. Instead, manufacturers must continuously manage consumption and cost. It requires a very different skillset as well as new operational functions. We recommend that manufacturers adopting cloud establish a cloud Center of Excellence (CoE) that works as a single focal point for both OT and IT. This can significantly accelerate cloud adoption and the associated value realization. The CoE brings central governance and direction to cloud architecture and design choices. This helps manage the complexities of distributed and multi-cloud solutions. And it prevents the confusion that can ensue if individual parts of the manufacturing organization go in differing directions.
Great advice … thanks for getting us up to speed on this.
No problem. And if you want to know more, don’t forget to check out our most recent report on this topic.
Thiago is Managing Director and the Industry X Manufacturing Execution Systems Lead in North America at Accenture. Start a conversation with him via LinkedIn.
<<< Start >>>