RESTK

RESTK is an implementation of the algorithm to upper-bound extremely high quantiles using Markov's inequality to the power-of-k for probabilistic WCET.

Institution:

Institution

Research Group:

BSC Group: Computer Sciences

Researcher/s:

Sergi Vilardell

Description:

RESTK is a statistical tool based on Markov’s Inequality that estimates very reliable worst-case execution times from measured data. By analyzing execution-time samples, it provides guarantees that deadlines will be met with extremely high confidence. This is essential for safety-critical systems in fields like aviation, automotive, and healthcare. The tool works with data from any platform and runs in R, making it broadly usable. More accurate timing estimates can also allow systems to run with fewer resources, indirectly improving energy efficiency.

Value Proposition:

Statistically tight WCET bounds for safer, leaner real-time systems

Aplication areas:

Autonomous-vehicle ECUs, avionics flight computers, medical imaging devices, industrial robots, space on-board computers, mixed-criticality drones

Target market:

Automotive Tier-1s & OEMs seeking ISO 26262 compliance; Avionics integrators targeting DO-178C / CAST-32A; Medical-device manufacturers under IEC 62304; Certification & testing labs, safety-analysis consultancies; Academia researching probabilistic timing analysis

Protection:

GPL License (Version 3.0)

More information

if you want to know more about this project do not hesitate to contact us

Contact us