Software as a service (SaaS) and infrastructure as a service (IaaS) provide services at opposite ends of the requirements spectrum. SaaS is something used for purpose-centric web applications like Gmail, Spotify, QuickBooks Online and PayPal. It offers ease of use for end users to get things done. Meanwhile, IaaS allows developers to run virtual machines (VMs) of their choosing, making it simple to spin up a Linux or Windows VM on a hosted virtual computer with user-specified memory, storage and number of cores. IaaS offers fully flexible systems for developers to create do-it-yourself solutions.
A paradigm that has been slower to emerge is found in the area between SaaS and IaaS—advanced configuration of software using platform as a service (PaaS). On our local machines, an Excel expert can automate worksheets with macros by turning on developer mode and running Visual Basic for Applications (VBA). These macros can be shared with peers through file copies.
Google Sheets is a cloud-based version that automates the use of Google Apps Script (JavaScript routines within Google Apps). Part of an app platform—or PaaS—these scripts allow creation of new data structures, reading of values from external sources, writing of custom reports, manipulation of values in the spreadsheets, sharing of results with others and a wide range of other advanced functions all by logging into a Google account. It is not necessary to know the operating system, information technology or networking details.
PLC configuration ready for PaaS
The desktop configuration environment for programmable logic controllers (PLCs) is being abandoned in favor of server-based configuration platforms. The infrastructure component of an automation platform is currently mired in complexity. When it comes time to create a secure server that hosts the operator screens, databases and analytics engines that allow these devices to be abundantly more beneficial, a wide range of questions must be answered: Where will it be hosted? What operating system (OS) will be used? How will it communicate? What drivers are needed? Just to decouple the computer with a virtualized solution means first creating custom images and setting all the security settings. IT teams need to design a network to load the configuration software and load the VMs so that it is possible to log on to a server-based PaaS.
Imagine powering up an automation server image—either running in your building or hosted by someone else—without knowing or caring what OS it is running, or how the databases are set up for history services/screen development/recipe management. Just log into the automation PaaS and start to configure your system. Furthermore, with appropriate authentication, you can query automation values, access reports and even supply parameters that the automation PaaS passes on to the control algorithms. This direction makes complete sense because the corporate infrastructure does not need to dictate the automation infrastructure or vice versa.
Business analytics is the golden ring
Many control engineers aren’t really that concerned with the infrastructure of the development environment. Most of the time, the development is targeted for a single installation site or copies of it for multiple sites. Where things really start to matter is at the analytics level. Drawing key performance indicators (KPIs), costing trends and performance matrices in a corporate environment adds a new level of perspective. The operating system or automation platform varies across lines or plants. At minimum, the version level of the automation platform will vary. What has emerged is the concept of “connectors” that bring the data from historians to the business analytics engines. The advantage is that historians store data in standardized ways that do not depend on the control platform from which they emerge.
The Everest Group has a study examining the analytics business process services focused on their ability to deliver capability against their company size. The leaders in this space are IaaS suppliers that are now packaging solutions the build a platform for business to perform analytics. This evolution further compels companies to have PaaS for the automation just as they have their analytics in PaaS.
Moving forward
What will it take for PaaS to emerge in automation and analytics? Just as in the Internet environment, automation suppliers must embrace the concept and create the platform. This is certainly an investment, but also achievable given the VM approach already used by many equipment suppliers. The other challenge will be support. System problems will fall into the laps of the PaaS provider, but there will only be one configuration to support.
There are also opportunities to simplify and capitalize on a more extensible licensing model of the PaaS—fully developed by the automation supplier in cooperation with the large IaaS suppliers that are building analytics PaaS engines. It seems like a win-win.
Mike Triassi is business development manager at Optimation Technology Inc., a certified member of the Control System Integrators Association (CSIA). For more information about Optimation, visit its profile on the Industrial Automation Exchange.
