After a recent customer visit with a market-leading OEM specializing in automation equipment for high-end printing operations, I’m more excited than ever about the potential of multi-core in embedded applications. This OEM specializes in designing Finishing Control Systems (FCS), which are installed on state-of-the-art equipment found in printing centers, such as binders, saddle stitchers, multi-mailers, poly-wrappers, envelope inserters, and presses.
Historically, the FCS has consisted of two distinct computers. One computer is dedicated to hard real-time automation applications—machine tracking, I/O control, record file handling and data collection-- that must be carried out in strictly deterministic fashion, only achievable within a hard real-time operating system framework. The second is a separate general purpose system which handles all non-deterministic applications in the enterprise domain---such as the user interface--built on commercial general-purpose operating framework. Currently, each of these two computers runs on an entirely separate board.
For their next generation of industrial automation platforms, the customer’s systems architects reached to the latest multi-core processors—the Intel Core 2 Duo. Beside the obvious cost/performance/power advantages of multi-core, they were attracted to the possibility of consolidating both—the real-time operating environment and the general-purpose applications—onto a single board/chassis. And although the processing power was there to make such evolutionary leap, there were formidable challenges. How to efficiently marry the two different operating environments to run side-by-side on a single chip, while at the same time:
- Guaranteeing hard real-time performance of their automation applications?
- Enabling communication between the two systems without impacting safety, security and performance of either?
- Efficiently migrating the thousands of man-hours invested into their proprietary code for industrial applications to the new multi-core system?
As it happens, a solution meeting these requirements is already available. The RadiSys Procelerant IS 1000 application-ready platform is based on Microware Hypervisor virtualization technology. This “embedded” hypervisor enables multiple operating systems to simultaneously run either multiple instances of a real-time operating system, or a heterogeneous mixture of operating systems on a single execution platform. Unlike hypervisor architectures common in the enterprise domain, the real-time operating system runs unmodified in assigned protected partitions that provide direct hardware access. Each operating system runs independently and at full speed, accessing assigned I/O devices without interference from the other OS, while sharing data and coordinating processes through shared memory and efficient virtual networks.
The virtualized environment of the Procelerant IS 1000 platform delivers on all fronts, as it:
- Guarantees real-time performance of industrial automation applications,
- Partitions real-time and general-purpose environments for enhanced safety, security and system peformance,
- Provides an easy migration path of proprietary application code to a multi-core platform.
Starting with a COTS application-ready platform like the Procelerant IS 1000 gives the OEM developers access to open, industry-standard technologies, including processors, boards, chassis, and software stacks such as industrial Ethernet and TCP/IP. Hardened for industrial applications, the Procelerant IS 1000 is already certified for reliable operations within the extreme conditionns of the industrial setting, such as extended temperature range of -25C to +70C. This complete framework allows application developers to focus on their value-ad – bringing innovative automation solutions to the market faster, with less risk and at lower cost.
And their customers will now benefit from the full potential of multi-core. With fewer computers in the virtualized system, their printing facilities will realize improved level of productivity, while achieving increased reliability (MTBF) of the system, diving better economics of the total cost of ownership.
What is your multicore story?
Director, Software & Integrated Systems