| KEYWORD |
Virtual Tool-boxing for Robust Management of Cross-layer Heterogeneity in Complex Cyber-physical Systems
keywords DIGITAL SYSTEM DESIGN TEST AND VERIFICATION, DOCKER, RELIABILITY
Reference persons STEFANO DI CARLO, ALESSANDRO SAVINO
External reference persons Alessandro Vallero, Alberto Carelli
Research Groups TESTGROUP - TESTGROUP
Thesis type RESEARCH
Description This thesis aims at developing tools able to decoupling the software and hardware development through virtualization while providing adaptability to heterogeneity, reliability and security to help make future Cyber-Physical Systems become a reality. The goal is to deliver an integrated development-production environment composed of a new container-based secure and resilient virtualization technology to deal with the complexity of the multi-dimensional heterogeneity (devices, systems, workloads, operating conditions, variability) plus a test/run-time set of monitorization, analysis and reconfiguration plugins to the virtual platform for enhanced energy-efficiency, life-time resiliency, security and updateability of future Cyber-Physical systems . The technologies developed in this thesis will cross-over all system stages from design to production; consequently, it will reduce the time-to- market as it enables a common hardware interface that eliminates the transition time between the design/prototype-test/ production phases. Virtualization will provide a degree of freedom unachievable with a conventional setup, plus low-overhead containers provide a feasible solution for CPS. Consequently, the new environment will reduce the system design cycle time by achieving the following:
• A virtual platform that can seamlessly be executed on simulators/emulators/final silicon. Thus, reducing development/ validation time and increasing productivity.
• A new paradigm to develop cross-platform toolboxes for characterization/testing/reconfiguration solutions.
• Transparent Low-level management of heterogeneity managed by the container engine.
• Transparent machine learning based per-application adaptability to Quality of Service (Qos) in terms of power, performance, timing, reliability and security through containerization .
Deadline 28/10/2021
PROPONI LA TUA CANDIDATURA