Memory Management allows relocation of Real-Time Operating Systems without virtualisation overhead and eliminates jitter on NUMA based hardware. By means of hardware based virtualisation, Version R4.3 of the RTS Hypervisor provides its users now with the possibility to freely relocate Guest Operating Systems in memory. This makes it possible to deploy 32-bit operating systems in memory above the addressable limit of 4 GB or to load and run multiple kernels linked to the same physical address in parallel without any virtualisation overhead, deterministically and in hard real-time. A benefit for users of high-end systems, running the RTS Hypervisor on multi-socket XEON platforms found, for example, in medical MRT machines, is the newly added support of the Non-Uniform Memory Access (NUMA). In the NUMA architecture, where each physical processor (NUMA Node) may have memory connected to its own local memory controller, the user can now assign memory of specific NUMA Nodes exclusively to an operating system. This greatly reduces memory access times and jitter, as simultaneous, competing memory access by multiple operating systems is avoided. With Version R4.3, Microsoft Windows 10 and Concurrent Computer Corporation’s RedHawk 7 is now supported by the RTS Hypervisor. Together with recently-added security features including access rights for shared memory partitions or the new rights management for the use of programming interfaces (APIs) to start, stop or otherwise control guest operating systems, the latest RTS Hypervisor platform can be applied to the Internet of Things (IoT), Industry 4.0, Test and Measurement or Medical applications, and for all embedded systems where determinism, hard real time or security is crucial.
The RTS Hypervisor R4.3 supports all current Intel x86 Multi-Core designs. Out of the box, the guest operating systems Microsoft Windows, QNX Neutrino RTOS, Wind River VxWorks, Linux including RedHawk, Windows Embedded Compact, Microware OS-9, On Time RTOS-32 and T-Kernel are supported.
Real-Time Systems; www.real-time-systems.com