Sramana Mitra: If you could, walk us through a couple of use cases where these concepts are playing out in your domain.
Steve Pavlovsky: That sounds fantastic. The core product we make is a control system that enables a user to control a machine. In effect, it’s a special purpose computer. These computers have typically Intel-style CPUs. There’s memory on them. They don’t run a traditional operating system like Linux or Windows. In our case, our systems run VX Works. They run a real-time operating system, and they have a kernel on them that processes a control algorithm to connect to machines and control the modes of a machine and what that machine does at any given time.
Traditionally, we’ve had programming software that runs on a laptop. In the old days it connected serially to these controllers, and then over Ethernet over the last 10 years. We see a paradigm shift from that laptop programming tool connected to a controller to a more systematic view, which is a system of controllers connected via Ethernet technologies and a network of users who are interacting with that network of controllers. The way we’re headed is to use the cloud computing technology platform to enable that interaction.
If you think about some of the paradigms we’re trying to change, in the real industrial world, one of these controllers controls a piece of machinery. It’s got a program in it that controls that machinery. In the middle of the night, something on the machine may break, and a maintenance person might make a small change to that program. He knows that. It’s on his laptop, but when somebody comes in to work the next day, the engineer responsible for that manufacturing plant might not know that that program has been changed. The cloud solves this problem in that if we store all of the programs for all of the control devices in the cloud, every user within that organization will always have access to the most up-to-date data.
It takes several engineers to program a complicated machine. The cloud computing platform will allow multiple users to work simultaneously to build more complicated applications.
SM: Is the assumption that the software is self-repairing by drawing resources from the cloud?
SP: Typically what the infrastructure would do was alert a human to a problem and enable that human to interact with the machine. Because we’re talking about physical machines that often move, if they move incorrectly, they could cause damage to themselves or could hurt somebody. There’s typically a human involved with making those kinds of changes.
Think about a plant. It has an engineer who’s home in bed. His pager or cell phone goes off, letting him know there’s a problem. He can get on his Web-enabled device and interact with the machine from his bedside as opposed to driving to the plant. That’s all new capability, that cloud-based infrastructure.
SM: The assumption is that the problem is in the software. If it’s a mechanical problem, somebody has to go in and fix it.
SP: Well, it may be a mechanical problem, but with the software, they can make a temporary fix. Perhaps a sensor has broken and they say, “OK, we’ll write some logic to bypass that sensor,” and that’ll be OK for the next eight hours. Those are the kinds of changes they’ll make. Or they could just dispatch a maintenance work order to have somebody who’s in the plant to go change the sensors. Those are the other kinds of things that could happen.
SM: That’s your regular workflow.
SP: Right. The cloud enables workers to do it from home or from around the world if necessary.