Products

Our Products

Dependable computing solutions

We develop cutting-edge fault-tolerant computing solutions to enhance the reliability and resilience of mission-critical systems in Aerospace, Automotive, HPC & AI, and Industrial Safety.

These tools are designed to evaluate, analyse, and test the reliability of computing systems, particularly SoC-FPGAs under radiation and fault conditions.

Fault Tolerance & Dependability Solutions

This category includes solutions that enhance system reliability, ensuring that computing systems can detect, recover from, and prevent faults.

Dependable computing for mission-critical applications

Ensuring high reliability in computing systems is crucial for space, aerospace, automotive, and industrial applications. Radiation effects, hardware faults, and system failures can compromise mission success.

Functional verification EDA (electronic design automation) tools SK-RAFT – emulation-based fault injection for functional verification

SK-RAFT is our flagship fault injection platform, designed for the functional verification of computing systems used in critical applications. It currently supports FPGA-based platforms, with full compatibility across all AMD-Xilinx devices. We are actively extending its capabilities to support a broader range of computing architectures, including:

  • FPGAs (supported)
  • Processors (extending)
  • GPUs (extending)
  • AI accelerators (extending)

Support for Arm-based platforms is planned for upcoming releases, enabling full-stack dependability verification in mixed FPGA-CPU environments in AMD Xilinx SoC FPGAs.

Verification: SK-RAFT enables emulation-based fault injection in alignment with key industry standards such as:

  • ECSS-Q-ST-60-13C (Space)
  • ISO 26262 (Automotive)
  • DO-254 / DO-178C (Avionics)
  • IEC 61508 (Industrial Safety)
  • IEC 62304 / IEC 60601 (Medical Devices)
  • IEEE 2851 (Generic fault injection guidance)

Key capabilities

Wide FPGA support

Currently supports fault injection in all Xilinx FPGA families, including 7-Series, Zynq-7000, Zynq UltraScale+, and Versal devices.

Analyses the FPGA’s essential configuration bits and restricts fault injection to these bits, improving test efficiency.

  • Stuck-at faults (0 or 1)
  • Single event upsets (SEUs)
  • Multiple bit upsets (MBUs)
  • Multiple cell ipsets (MCUs)
  • Custom fault scenarios

Faults are injected dynamically through JTAG into the FPGA while the circuit runs, enabling noninvasive, real-time evaluation of system response.

Supports Virtual Input/Output (I/O) using the BSCAN primitive inside Xilinx FPGAs, allowing the user to stimulate and observe I/Os without altering the user design.

SK-RAFT is compatible with USB-connected JTAG FTDI programmers and integrates seamlessly with popular FPGA development boards.

Use cases

  • Verifying FDIR strategies in safety- and mission-critical systems
  • Evaluating SEE vulnerability in FPGA applications 
  • Functional verification for automotive ADAS or avionics subsystems
  • Supporting standards compliance through systematic fault injection campaigns
  • Monitoring FPGA applications during accelerated radiation tests

FPGA IP cores for fault tolerance

SK-FDIR is Solidkosmos’ growing library of software and hardware IP cores designed to increase the fault tolerance of FPGA-based systems in critical applications. It provides various essential software and hardware (VHDL) modules to implement Fault Detection, Isolation, and Recovery (FDIR) strategies in FPGAs. The library is optimised for AMD-Xilinx FPGA-enabled embedded platforms and is validated through real radiation testing.

The SK-FDIR library includes the following products

SK-RECOV – Error Recovery for Lockstep Execution

Developed under an ESA OSIP framework, SK-RECOV introduces advanced fault recovery capabilities for the dual-core Arm A9 CPUs in Zynq FPGAS

Compared to state-of-the-art, SK-RECOV improves:

  • Fault coverage
    • Can detect errors in the FreeRTOS kernel space
  • Common-mode errors
    • Maps checkpoint data to (isolated) physically separate memories
  • Performance overhead
    • CPU caches are enabled
    • Checkpoint & Restore is IP implemented in the PL 

Results

  • Performance overhead: 0.75% to 7.5% for SysTick 10-100 ms
  • Area overhead: TMR IP core on a Zynq-7020: 14% LUTS, 5% Regs
  • Neutron Radiation Tests: 83% success recovery rate,  213 of the 260 radiation-induced errors were recovered successfully during a 102-minute accelerated radiation campaign. 



SCRUB is a lightweight, hardware-assisted memory scrubbing engine that protects FPGA configuration memory from Multiple Bit Upsets (MBUs) with minimal resource requirements.

  • Requires just 7.1 Mbit of flash memory, compared to 128 Mbit used by conventional solutions (e.g. AMD SEM Controller with frame replacement) to correct MBUs
  • Supports both:
    • External scrubber (SW-based, runs on external rad-hard CPU)
    • Internal scrubber (HW-based, instantiable in the PL)
  • Compatible with Xilinx 7-Series, Zynq-7000, and Zynq UltraScale+ devices
  • Validated through heavy-ion radiation testing

SK-MON is a C/C++ software library that runs on embedded or external CPUs to provide real-time health surveillance of SoC-FPGAs.

  • Logs critical runtime metrics such as:
    • RPU/APU performance
    • Voltage and temperature fluctuations
    • Error rates
  • Includes diagnostic algorithms to predict system failures
  • Offers customisable alerts based on user-defined thresholds

Discover the science behind our solutions

See how our fault-tolerant computing solutions power high-reliability systems across Aerospace, Automotive, AI, and Industrial applications.