This Small Business Innovation Research (SBIR) Phase I project proposes the creation of a mixed-criticality Java virtual machine (VM) and native code hypervisor of unprecedented flexibility and capability. Current generation mixed-criticality systems rely on proprietary and expensive operating systems, limiting the adoption of such systems and their attendant benefits. Classic mixed-criticality approaches trade off expressivity and flexibility for safety and verifiability. As such, they are currently limited to specialized hard real-time market sectors, like avionics. Cutting edge research in compiler design, virtual machine scheduling, transactional memory, and mixed-criticality communication will be combined with an existing VM and traditional mixed-criticality hypervisor to create a nontraditional mixed-criticality hypervisor combining the best features of hard real-time Java, mixed-criticality systems, and flexible off-the-shelf software development capabilities. This research will address just in time compilation and runtime code insertion for hard real-time environments, flexible soft real-time time partitioning for virtual machine hypervisors, hard real time transactional memory protocols, and time- and space-partitioned hard real time communication mechanisms. The broader impact/commercial potential of this product is the great reduction of time and cost for development and increased reliability of mixed-criticality software systems. Embedded real-time software can be found in automobiles, aircraft, satellites, medical devices, and many other applications. Java-based mixed-criticality solutions offer to reduce development time and costs, while attaining high fidelity software with provable guarantees. By reducing time-to-market and costs, the technology in this proposal will enable faster innovation in a potentially large number of embedded software markets. More specifically, the cost-effective creation of the next generation of medical devices with provable guarantees ensuring both patient health and patient privacy, thereby enabling the further modernization of medicine, while adhering to regulatory guidelines set forth by government entities. In addition, the software developed in this proposal will be available free of charge as open-source software to be used by the academic and scientific communities for non-commercial purposes
