Faced with so many data analytics options—and high prices for failing—it can be welcome to have some guidance.
“Where they previously stayed mostly at the manufacturing execution systems (MES) layer, many clients are now going to cloud-computing services for analytics, so we’re working with more software suppliers and information technology (IT) departments to coordinate specifications, requirements, features and functions,” says Pradeep Paul, manufacturing intelligence director in the IT division at E Tech Group, a system integrator in West Chester Township, Ohio, and certified member of the Control System Integrators Association. “For example, when they employ or plan to offer software as a service (SaaS), it can give us data, too. Previously, we’d install a reporting package, but now these capabilities require much less customization.”
Get your subscription to Control's tri-weekly newsletter.
Just as it matches end-users’ operational requirements with suppliers’ suitable hardware and networking, E Tech works with clients to determine what features their SaaS needs to perform. However, because there are so many SaaS providers, it can be hard for any single user to sufficiently vet them all. “We bridge that gap. We have use cases and demos that can put clients ahead in the SaaS evaluation game,” says Paul. “This lets them review different suppliers and their capabilities to acquire suitable services.”
Escape from clunky
Even though different mission-critical applications employ various equipment and software in different arrangements, Nick Hasselbeck, VP of IT in E Tech’s IT group, reports that projects at multi-megawatt data centers or building-management systems still share some common threads and strategies as they seek to optimize or upgrade.
“Many applications begin with a SQL database, and apply classic, web-based Ignition SCADA/HMI software to monitor energy consumption,” says Hasselbeck. “Our group’s engineers will then assemble a Microsoft Azure application to process time-series information about local energy use, and locate it in a central area that authorized users can access, including exposing to the company’s management and business levels.”
Similarly, E Tech constructs applications for its consumer-packaged goods (CPG) clients consisting of a customized middleware layer that links enterprise resource planning (ERP) software to data from downstream production lines. “This connection lets users assign or adjust recipes from the enterprise level, manage changeover processes manage recipes, and send related documents to PLCs or other devices affected by the changeover,” explains Cole Switzer, advanced software engineering manager in E Tech’s IT group. “In the past, these tasks were mostly manual or just had some preprogrammed setpoints.”
Kevin Romer, solutions architect in E Tech’s IT group, confirms that former, dedicated applications relied on early, physical servers with typically rigid, non-virtual functions. “They sent instructions to PLCs via a SQL server,” says Romer. “They ran custom applications onsite, along with third-party software, such as iFix or Wonderware, but they weren’t very flexible.”
Paul reports that Ignition’s digitalized, web-based SCADA functions make it easier to deploy, configure and interact with data from batch software and DCSs, such as Emerson’s DeltaV and Rockwell’s PlantPAx. “Because Ignition is browser-based, it lets the Internet of Things (IoT) participate in a system-agnostic SaaS that can use HTML 5 to run on tablet PCs, smart phones and other mobile devices,” explains Paul. “This means users are no longer stuck at desks or other fixed locations. More specifically, it used to take six weeks to changeover a typical CPG production line to accommodate different infeed materials or produce different products, but now that time is down to one week.”
Romer adds that quicker changeovers means users no longer have to produce as much as possible of one product, and store it before the lengthy preparation for making another product. Faster transitions also enable a company’s data center to handle explosive growth in demands on its applications and businesses. “Older data architectures and solutions can’t handle these changes. One button click would be followed by 30 seconds of churning, while billing requests could take weeks,” says Switzer. “Now, responses from web-based software and tools are immediate, and users get near real-time data on power consumption and other parameters, which allows faster and more accurate billing.”
Slice, dice and containerize
While web-based software can help analytics, Switzer adds they’re also aided by how each organization’s data is partitioned.
“Most users previously had one huge database in a horizontal architecture, but the recommendation now is to break it into hundreds of parts based on software containers. “For example, our CPG client wanted to transfer their files from one monolithic database with a single application program interface (API) to a RESTful API and one database at each of 12 buildings or 1,000 sites,” explains Switzer. “We put an Azure gateway in front, so the client could view our panel access, and monitor all of the small applications in those 1,000 databases in their horizontal, distributed architecture. Because one database on one server is limited, it’s more flexible to spin up as many databases and distributed applications as needed.”
Switzer reports that containerized applications are also more efficient because they’re built on common software types, which are already familiar to many users, such as like Representational State Transfer (RESTful), Simple Object Access Protocol (SOAP), JavaScript Object Notation (JSON), Message Queuing Telemetry Transport (MQTT) publish-subscribe protocol, and NodeRED flow-based, low-code development software. “Containers are to virtual machines (VM) what VMs are to physical servers,” adds Switzer. “This is why it helps to use common development tools and protocols, which can all talk to each other, and perform tasks that can be scaled up.”
Hasselbeck explains that monitoring power consumption or other performance parameters occurs on a hybrid platform, such as a virtual server or cloud-computing service. However, sensors, cables and physical servers are usually on-premises, which is why they require a gateway like the Azure version E Tech uses to reach a SCADA/HMI stack and RESTful API in the cloud, allowing users to benefit from the best aspects of each side. E Tech refers to its bridging capability as a “product as a service” (PaaS) or customer-specific SaaS.
“This can alleviate bottlenecks in processes by taking thousands of formerly separate, individual reports, and turning them into one centrally managed reporting application,” says Hasselbeck. “It allows users to have better interactions with their data, which can also be finer-grained at the millisecond-level, rather than the tenths of seconds or full seconds they experienced before.”
For example, E Tech’s prior work with its CPG and data center clients gave them to trust to team-up on new opportunities, such as a middleware solution and custom-cloud application connecting their ERP to various downstream processes via Kepware drivers distributed to mechanical components. In fact, the CPG client has deployed this solution at about 100 plants.