“Cloud is about how you do computing, not where you do computing” Paul Martiz, CEO VMware
Cloud computing is seen as a formidable driving force from an Industry 4.0 perspective, alongside other enabling IT technologies such as simulation, analytics, IoT and mobility, which joined together to form the basis for this new industrial revolution. It is estimated that the cloud industry is expected to grow from $371 billion in 2020 to $832 billion in 2025—a staggering CAGR (compound annual growth rate) of 17.5%.
When it comes to the role of the cloud in manufacturing and how MES as an application deals and adapts to it, it is still a fairly nascent practice; there is no standard methodology to be followed or established best practices. What is established is the cloud framework—the hosting options—with various launch points from there.
Today, we will look at the foundational aspects of MES and the cloud to understand what cloud computing offers, what makes it so exciting for businesses around the world to adopt for manufacturing, and how the MES interplays and leverages the cloud. We will review possibilities without prescribing specifics, as the extent to which the role of cloud computing extends in a particular value chain requires granular analysis of the operation both on the shop floor and in IT, along with the needs and priorities of the process owners. There is also the obvious consideration of the capability of the existing applications and infrastructure to either adopt or be replaced by native cloud-based applications.
The basics: what is the cloud?
But first, let’s understand cloud computing a little better, starting with a basic definition of what comprises a ‘cloud.’ The cloud can be described as a secure, elastic, pooled platform with three types of services or offerings for hosting: Infrastructure as a Service; Platform as a Service and Software as a Service.
As you can see from the image below, the offerings vary by the amount of content hosted in the cloud, versus hosted outside of the cloud architecture: on premises being 100% physical, while SaaS (Software as a Service) is 100% virtual (that is, 100% based in the cloud).
In addition to the cloud offering types (above) there are three deployment methods: Private, Hybrid and Public.
Let’s take a closer look at each offering.
1) On premises cloud deployment/application hosting. This is referred to as a ‘private’ cloud. There is a single user (tenant) and the cloud is not shared outside of the company. This option is viable for organizations concerned about security, IP protection or application performance (i.e., real-time analytics and decision making).
2) Hybrid cloud deployment. The hybrid cloud is a mix of private and public clouds. This configuration makes sense where sensitive IP and application functionality might be required locally, but data back-up and storage of the raw data itself might be on public (multi-tenant) cloud.
3) Public cloud deployment. The public cloud uses off-site cloud deployment and hosting. The application, along with its database and functionality (such as backup and replication) fully resides on a public cloud network. These are called multi-tenant deployments– companies share cloud storage, and their data resides alongside other companies. From a cost perspective, a public cloud is normally the least expensive option, and is used for scalability and accessibility.
Most manufacturing companies use a mix of private and hybrid, as their operational needs (manufacturing data/IP protection, response times, scalability and reliability/uptime) require a less-risky, more controlled option.
As a client, choosing one architecture over another is a matter of preference as well as performance—your needs for security, reliability, access and scalability drive your decisions. For MES, since it has both real-time components (data acquisition) as well as near-real time components (reporting and analysis) it typically uses a blend of the cloud architectures for application hosting and data management.
Additional benefits of cloud hosting
The ‘beauty’ of a cloud-based service is the ability for incremental deployment, owning to its inherent scalability by design. It allows organizations to convert IT-related expenditures, which would traditionally be aligned with a capital expenditure budget, to an Operations expenditure, avoiding unnecessary delays for implementation.
Other benefits for the cloud have to do with deployment: reducing costs from infrastructure, hardware investment, and IT overhead.
- It allows remote hosting and deployment of IT applications
- Through pooled resources and infrastructure over the internet, it provides data security through redundant storage allowing instances of data loss to be eliminated
- It reduces upfront expenditures either through metered billing or based on the extent of an application being used, or a mixture of both models
Irrespective of the configuration being used, the appeal that cloud has to offer becomes apparent from the very features any cloud deployment possesses or rather promises to possess. It can directly impact the cost of IT related infrastructure, as it may eliminate the need to have on premise infrastructure for hosting applications and storing data; it impacts the cost of IT application in use as well, as users need pay only for the functionality they use, when they use it and for the data they store in the cloud for the time period they wish to store it. Data on the cloud is secure, with most cloud providers offering double or triple redundancy and backup. Lastly, uptime and latency isn’t generally an issue unless there is a major failure which disrupts the internet or connection.
MES and the Cloud
Industry 4.0 as a pre-requisite demands end-to-end digitization of a given value chain. If the benefits of cloud computing are realized and tailored for said value chain, it is a strong contributor in making the desired digitization possible.
One aspect of digitization is the large-scale data collection throughout the entire value chain—with manufacturing’s real-time aspects (through edge/IIoT) being an important contribution to the overall data pool.
MES is already well-known as an enabler of Industry 4.0. The distributed intelligence of a Modern MES inherently accommodates the Internet of Things, mobile devices, automation, and sensors. It offers advanced analytics, continuous improvement tools and performance enhancements such as digital thread/digital twin, AR and VR.
There are a few possible ways in which the Modern MES and cloud computing can come together: application hosting and data storage/replication. The choices you make should be influenced by your vendor selection (do they even offer cloud versions); your IT infrastructure (are you cloud ready) and your organizational strategy (will it aid Industry 4.0 progress, fit your deployment strategies and suit your operational needs).
An MES application may fully reside on the cloud, which implies that the entire application along with its functionality, data storage, data processing and data backup reside entirely on the cloud. This is a typical ‘SaaS’ or Software as a Service configuration. In such a scenario, only the client side is on premise. The choice of the cloud host itself (and associated costs) may vary depending on tactical features such as uptime, transaction volume, and desired data integrity (replication), as well as more strategic decisions as to data/IP security, and overall performance of the application. The closer to real-time performance desired, the more likely an on premises hosting is chosen. Of course, if you already have applications on the cloud, the decision is much more straightforward. Then, it’s a matter of how your MES application provider supports your specific cloud architecture.
In the scenario where MES fully resides on the cloud, application deployment and costs related to cloud deployment may be lower than full scale traditional (on-site) deployment, but there may be concerns pertaining to latency, lack of real-time performance and data security. An alternate arrangement may then be to have vital MES functions hosted either on-site or on a private cloud, thereby allowing for better control and security as opposed to a fully public/cloud-based application deployment. Functions like process execution, scheduling, dispatch, data collection and data analysis (especially at the edge, closer to the control layer), might be kept local and not hosted on the cloud, while other functionalities, such as quality control, R&D, LIMS, compliance and/or safety data may be hosted on the cloud. Such a setup ensures that the vital aspects of the MES are always up and that data which is considered extremely critical is within the organization’s perceived control.
After considering various/possible MES hosting and cloud configurations, the question that comes up is obviously, which is the best possible configuration to use? The answer to this question is not obvious though: it depends on the organization’s perspective on data and application up-time, security, network latency, costs of deployment, scale of deployment and possible control on data and applications. The way in which the MES utilizes cloud computing may vary, which brings us back to the quote with which we started, it is not just where you compute rather how you compute that defines cloud.
The choice is yours
MES vendors may offer fully or partially cloud-enabled applications. The important take away is that the decision to use the cloud should be based on your organization’s IT infrastructure, coupled with manufacturing’s process needs. Do you have a single site, or multiple sites that the MES must manage? Are you already on an Industry 4.0/digital transformation journey (meaning, key decisions have been made on technology, architecture, performance and stakeholders) or is this a blank slate? Choose an MES that can deliver on these needs or use cases. The cloud is a delivery mechanism that helps you reach your Industry 4.0 goals; it is part of your solution set to gain agility, efficiency and optimization. If this can be delivered incorporating the cloud and its benefits—excellent; if not, then the need to be more digitized and deliver quantum improvements on key use cases should take top priority.
Join the Manufacturing Data Revolution