Why do most internet of things (IoT) analytics operations occur in the cloud? The public cloud offers a centralised location for large amounts of affordable storage and computing power. But there are many instances in which it makes more sense to perform analytics closer to the thing or activity that is generating or collecting data – equipment deployed at customer sites.
This is particularly true in industrial and manufacturing environments, which are familiar with the challenges of managing massive amounts of unstructured data, but may lag when it comes to the virtualisation of IT infrastructure.
Industrial cloud market development
Advances in intelligent process manufacturing, factory automation, artificial intelligence and machine learning models all benefit from edge analytics implementations, yet will likely become islands of automation without a cohesive industrial cloud computing platform.
The industrial cloud covers everything from the factory floor to the industrial campus, and it is unifying the supply chain as companies employ a combination of digital business, product, manufacturing, asset, and logistics planning to streamline operations across both internal and external processes.
Industrial cloud applications make it easier to optimise asset and process allocations by modeling the physical world and use data and subsequent insights to enable new services or improve control over environmental, health, and safety issues.
The virtualisation of business-critical infrastructure is transforming the production and distribution of goods and services throughout the supply chain, as industrial organisations shift focus to hybrid cloud computing deployments that connect and integrate on-premise IT resources with public cloud resources.
According to the latest worldwide market study by ABI Research, hybrid industrial cloud adoption will more than double over the next five years at a respectable 21.1 percent CAGR.
Initial IoT deployments in industrial markets reflect the sector's machine to machine (M2M) heritage: private cloud infrastructure as a service (IaaS). The private IaaS model served as a solid starting point for many organisations that wanted the benefit of cloud scale, but with minimal interruption to normal IT operations.
The industrial cloud platform as a service (PaaS) model extended the functional capabilities of on-premise IaaS solutions by shifting commodity tasks – such as capacity planning, software maintenance, patching — to public cloud service providers. Software as a service (SaaS) took it a step further but in the form of managed services.
"Manufacturing and industrial organisations were not born from the same digital core as the people they employ or the products they produce," said Ryan Martin, principal analyst at ABI Research. "But they also harness some of the greatest potential thanks to massive amounts of untapped plant and process log data. Harvested with the right analytical tools and guidance, these data streams can deliver value greater than the sum of their parts."
The factory floor’s historical predisposition toward on-premise solutions has been supplanted by a campus-led approach underscored by a more recent push to connect HMI, SCADA, and control networks to higher-level enterprise systems, as well as the public cloud.
However, getting to the point where all these moving pieces come together in a real-world, production environment can be messy. Many operational technology (OT) devices come up short in key areas such as interoperability and security due to the prevalence of proprietary protocols in the legacy M2M market.
Outlook for industrial cloud app development
"Most OT systems depend on infrastructure with lifetimes measured in decades, while IT systems can be upgraded frequently at little or no cost," concludes Martin.
As a result, industrial and manufacturing markets typically employ a staged technology integration strategy that favors suppliers whose hardware, software, and services can be acquired incrementally, with minimal disruption to existing operations. The hybrid IT infrastructure models can fit very well in this operational environment.