Search

International audience; This paper presents a framework and implementation guidelines to set up nested compartmentalisation in constrained devices. All memory spaces are protected by the Memory Protection Unit (MPU). Current MPU-based systems offer efficient memory protection but are mostly tied to the fixed permission model provided by their operating system, kernel, hypervisor or by code instrumentation. New use cases evolve with the rise of the Internet of Things (IoT) ecosystems where software components could benefit from locally and dynamically established permissions. This includes a temporary nested subspace with restricted memory access rights. Our framework integrates subspace creation and management for runtime dynamic changes of the permission model for any level of abstraction. Global security policies of fixed permission models are reflected in the software architecture and the implementation of the framework. We also demonstrate the feasibility of providing nested compartmentalisation by showing how to leverage the MPU features.

In mixed-criticality systems, applications naturally have different safety criticality levels. Partitioning technology is usually used to enable the integration of such mixed criticality applications upon one platform, aiming at reducing hardware, power consumption and especially certification cost. Partitioning can prevent fault propagation among mixed-criticality applications, if spatial and temporal isolation are adequately ensured. This paper focuses on the solution of spatial isolation in the HARTEX kernel on a multi-core platform in terms of memory, communication between applications and I/O sharing. According to formulated isolation requirements, a simple partitioning multi-core hardware architecture is proposed using SoC and memory protection units, and the kernel is extended to support spatial isolation between the kernel and applications as well as between applications. Combined design of hardware and software can easily achieve this isolation. At last, the spatial isolation is evaluated using a statistical sampling method and its performance is tested in terms of task switch, system call and footprint.

In the embedded systems domain, hypervisors are increasingly being adopted to guarantee timing isolation and appropriate hardware resource sharing among different software components. However, managing concurrent and parallel requests to shared hardware resources in a predictable way still represents an open issue. We argue that hypervisors can be an effective means to achieve an efficient and predictable arbitration of competing requests to shared devices in order to satisfy real-time requirements. As a representative example, we consider the case for mass storage (I/O) devices like Hard Disk Drives (HDD) and Solid State Disks (SSD), whose access times are orders of magnitude higher than those of central memory and CPU caches, therefore having a greater impact on overall task delays. We provide a comprehensive and up-to-date survey of the literature on I/O management within virtualized environments, focusing on software solutions proposed in the open source community, and discussing their main limitations in terms of realtime performance. Then, we discuss how the research in this subject may evolve in the future, highlighting the importance of techniques that are focused on scheduling not uniquely the processing bandwidth, but also the access to other important shared resources, like I/O devices.