Industrial manufacturers have recognized the
many benefits of plant floor digitalization.
Leaders in both IT and OT (operations technology)
must work together to decide on the degree
of information that needs to be collected as well
as what type of analytics are required. Due to the
wide range of requirements from other departments,
such as operations, marketing, finance, and
maintenance, pre-planning is the key for any IIoT
(Industrial Internet of Things) application.
New consumers of industrial data want immediate
opportunities to improve efficiency, profitability
or prevent downtime in ways that were
not possible before. Cloud services and scalable
storage are certainly drivers of this trend. Many
controllers, for example, support the mainstream
protocols commonly deployed in factories, but
also add new messaging schemes like MQTT and
REST APIs to support these new applications.
Cloud challenges
There are challenges with these approaches, for sure.
For example, cloud providers have their own set of
APIs and services unique to their solution. Customers
who are looking to adopt a cloud platform need to
consider how portable their data is and how easily they
can tie into existing systems within their organization.
A controller that allows for open development,
especially those that are Linux-based, will generally
support more platforms than a closed system.
Cloud platforms commonly use MQTT for
secure IIoT communications, leaving the integrator
to determine how the JSON payload is
structured; this adds flexibility but can also create
interoperability challenges. Specifications
like SparkPlug B help by defining the namespace
to address this challenge. But while Sparkplug B
is supported, relatively few cloud platforms have
formally adopted it yet.
Cloud platforms are constantly evolving. For
example, if you are using AWS Greengrass on a
PLC and AWS releases a new machine learning
service, a customer can easily adopt this new
feature. A new driver or update used to mean
installing and licensing new software on multiple
servers, hoping there are no conflicts. Fortunately,
this labor-intensive task might soon be a
thing of the past.
First steps
Data collection, storage and retrieval is the first
step in digital transformation. Many cloud agents
can orchestrate and manage your assets in a centralized
way, allowing changes to be scheduled
and rolled back as necessary. It is also possible to
remotely deploy code to a group of edge devices
for non-real-time tasks. This is a fundamental
change in the way controllers can be programmed
and managed, potentially making a significant
impact on our industry.
Central to the IIoT approach is edge computing—
collecting and processing data where it is
generated. There is a lot happening in the edge
of network space. This is an enabling technology
that greatly simplifies software development
and testing, which is more in line with mainstream
software development.
Adopting an open and easy approach to industrial
controls and supporting IIoT is accomplished
by designing controllers that support established
protocols and programming specifications, while
at the same time extending support for new solutions
or hybrid control schemes. This approach gives
controls engineers flexibility to deploy the software
applications that best meet their design goals.
