Q&A with Real-Time Systems
Communications are much faster since you are no longer communicating across cables. Instead, communication happens through a low-latency shared memory space with deterministic time synchronization of operations.
Tell us about Real-Time Systems (RTS) and the history of the company.
Fifteen years ago, when innovators began moving from single-core to affordable, higher-compute, lower-power multicore processors from Intel and AMD, our company’s founders realized that workload consolidation was possible. Working then as software architects, designers, and developers at KUKA Robotics, they founded Real-Time Systems (RTS) as a spin-off. They pursued their vision of matching multicore processors‘ capabilities to the market needs of robotics and industrial automation. Today companies realize the benefits of the RTS hypervisor, including real-time performance and reduced hardware costs.
What is it that sets your hypervisor apart from traditional hypervisors?
Real-time Systems’ hypervisor interacts directly with hardware enabling it to create partitions. This partitioning distinguishes the RTS hypervisor from others in the market, relying on virtualized hardware or emulators. Our competitors have guest operating systems and applications sitting on top of the hypervisor software stack. Even if you are on a bare-metal hypervisor, your applications are still running on top of emulated hardware, not the real hardware. Going through this software layer will add latency and jitter.
If you need real-time performance, you are not going to get it on emulated hardware. Our hypervisor delivers real-time performance.
To do this partitioning when hosting a real-time operating system, we take a multicore computer and first separate the CPU cores, memory, I/O, and PCIe devices. Then we allocate them safely and securely for each operating system.
How are hypervisors being used in robotics?
Rather than having a robot with separate hardware computers for systems such as HMI, vision, motion, data analytics or safety, you can have a robot that places many of these computer systems onto one computer. Our hypervisor makes robotic systems elegant and straightforward by assigning CPU cores, PCIe devices and SATA memory to each system where it runs in its own secure space. Furthermore, communications are much faster since you are no longer communicating across cables. Instead, communication happens through a low-latency shared memory space with deterministic time synchronization of operations. You can also use shared memory for exchanging data. And this is all taking place on one computer where the extra cabling, rack space and computers have vanished, saving substantially per robot system.
You are premium partners with Intel and support TSN and TCC with their new 11th generation processors, formerly known as Elkhart Lake and Tiger Lake. Why is that important and relevant in robotics?
It’s important and relevant for the advantages of partnering with Intel and what support for TSN and TCC offers for robotics. For robotics, you need reliable deterministic communications and performance with low latency and minimum jitter. A key component of achieving those goals is the tight integration of partitioning software with TSN and TCC. We have partnered with Intel since 2008, and during the last two years, we have been working closely with Intel’s Elkhart Lake and Tiger Lake engineering team to make that tight integration with our hypervisor possible. This effort brings leading-edge silicon capability to robotics with a hypervisor that will support the TSN network interfaces and TCC extensions.
Do you want to learn more about hypervisors in Robotics?