Search

Edge computing processes data near its source, reducing latency and enhancing security compared to traditional cloud computing while providing its benefits. This paper explores edge computing for migrating an existing safety-critical robotics use case from an onboard dedicated hardware solution. We propose an edge robotics architecture based on Linux, Docker containers, Kubernetes, and a local wireless area network based on the TTWiFi protocol. Inspired by previous work on real-time cloud, we complement the architecture with a resource management and orchestration layer to help Linux manage, and Kubernetes orchestrate the system-wide shared resources (e.g., caches, memory bandwidth, and network). Our architecture aims to ensure the fault-tolerant and predictable execution of robotic applications (e.g., path planning) on the edge while upper-bounding the end-to-end latency and ensuring the best possible quality of service without jeopardizing safety and security.

With the increasing use of multicore platforms to realize mixed-criticality systems, understanding the underlying shared resources, such as the memory hierarchy shared among cores, and achieving isolation between co-executing tasks running on the same platform with different criticality levels becomes relevant. In addition to safety considerations, a malicious entity can exploit shared resources to create timing attacks on critical applications. In this paper, we focus on understanding the shared DRAM dual in-line memory module and created a timing attack, that we named the "bank & row conflict bomb", to target a victim task in a multicore platform. We also created a "navigate" algorithm to understand how victim requests are managed by the Memory Controller and provide valuable inputs for designing the bank & row conflict bomb. We performed experimental tests on a 2nd Gen Intel Xeon Processor with an 8GB DDR4-2666 DRAM module to show that such an attack can produce a significant increase in the execution time of the victim task by about 150%, motivating the need for proper countermeasures to help ensure the safety and security of critical applications., Comment: To appear in the proceedings of the 27th IEEE international Symposium on Real-time Distributed Computing (ISORC)

Industries are considering the adoption of cloud and edge computing for real-time applications due to current improvements in network latencies and the advent of Fog and Edge computing. Current cloud paradigms are not designed for real-time applications, as they neither provide low latencies/jitter nor the guarantees and determinism required by real-time applications. Experts estimate that data centers use 1% of global electricity for powering the equipment, and in turn, for dealing with the produced heat. Hence, energy consumption is a crucial metric in cloud technologies. Applying energy conservation techniques is not straightforward due to the increased scheduling overheads and application execution times. Inspired by slot shifting, we propose an algorithm to support energy-aware time-triggered execution of periodic real-time VMs while still providing the ability to execute aperiodic real-time and best-effort VMs in the slack of the time-triggered ones. The algorithm considers energy reduction techniques based on dynamic power management and dynamic voltage and frequency scaling. We implement our algorithm as an extension to the Linux kernel scheduler (for use with the KVM hypervisor) and evaluate it on a server-grade Intel Xeon node., Comment: to appear in Real-time Cloud (RT-Cloud) 2023 workshop co-hosted with 35th Euromicro conference on Real-time systems (ECRTS)

There is increasing interest in using Linux in the real-time domain due to the emergence of cloud and edge computing, the need to decrease costs, and the growing number of complex functional and non-functional requirements of real-time applications. Linux presents a valuable opportunity as it has rich hardware support, an open-source development model, a well-established programming environment, and avoids vendor lock-in. Although Linux was initially developed as a general-purpose operating system, some real-time capabilities have been added to the kernel over many years to increase its predictability and reduce its scheduling latency. Unfortunately, Linux currently has no support for time-triggered (TT) scheduling, which is widely used in the safety-critical domain for its determinism, low run-time scheduling latency, and strong isolation properties. We present an enhancement of the Linux scheduler as a new low-overhead TT scheduling class to support offline table-driven scheduling of tasks on multicore Linux nodes. Inspired by the Slot shifting algorithm, we complement the new scheduling class with a low overhead slot shifting manager running on a non-time-triggered core to provide guaranteed execution time to real-time aperiodic tasks by using the slack of the time-triggered tasks and avoiding high-overhead table regeneration for adding new periodic tasks. Furthermore, we evaluate our implementation on server-grade hardware with Intel Xeon Scalable Processor., Comment: Accepted at the 17th annual workshop on Operating Systems Platforms for Embedded Real-Time applications (OSPERT) workshop 2023 co-hosted with 35th Euromicro conference on Real-time systems. OSPERT proceedings: https://www.ecrts.org/wp-content/uploads/2023/07/ospert23-proceedings.pdf

The Logical Execution Time (LET) model has been gaining industrial attention because of its timing and data-flow deterministic characteristics, which simplify the computation of end-to-end latencies of multi-rate cause-effect chains at the cost of pessimistic latencies. In this paper, we propose a novel method to reduce the pessimism in the latencies introduced by LET, while maintaining its determinism. We propose a schedule-aware LET model that shortens the lengths and repositions LET's communication intervals resulting in less pessimistic end-to-end latencies. By adding dependencies between specific task instances, the method can further reduce the pessimism in the latency calculations of the LET model. If needed, e.g., for legacy reasons, our method can be applied to a subset of tasks only. We evaluate our work based on real world automotive benchmarks and randomly generated synthetic task sets. We compare our results with previous work and the LET model. The experiments show significant reductions of worst-case data age and worst-case reaction latency values., Comment: Part of this paper was published at ERTS 2024-12th European Congress Embedded Real Time System. June 2024 - https://hal.science/hal-04617092/

Industries are considering the adoption of cloud computing for real-time applications due to current improvements in network latencies and the advent of Fog and Edge computing. To create an RT-cloud capable of hosting real-time applications, it is increasingly significant to improve the entire stack, including the containerization of applications, and their deployment and orchestration across nodes. However, state-of-the-art orchestrators (e.g., Kubernetes) and underlying container engines are designed for general-purpose applications. They ignore orchestration and management of shared resources (e.g. memory bandwidth, cache, shared interconnect) making them unsuitable for use with an RT-cloud. Taking inspiration from existing resource management architectures for multicore nodes, such as ACTORS, and for distributed mixed-criticality systems, such as the DREAMS, we propose a series of extensions in the way shared resources are orchestrated by Kubernetes and managed by the underlying Linux layers. Our approach allows fine-grained monitoring and allocation of low-level shared resources on nodes to provide better isolation to real-time containers and supports dynamic orchestration and balancing of containers across the nodes based on the availability and demand of shared resources., Comment: 6 pages, 2 figures, Accepted at the Real-time Cloud Workshop of the 34th Euromicro Conference on Real-time Systems

Industries are recently considering the adoption of cloud computing for hosting safety critical applications. However, the use of multicore processors usually adopted in the cloud introduces temporal anomalies due to contention for shared resources, such as the memory subsystem. In this paper we explore the potential of Intel's Memory Bandwidth Allocation (MBA) technology, available on Xeon Scalable processors. By adopting a systematic measurement approach on real hardware, we assess the indirect memory bandwidth limitation achievable by applying MBA delays, showing that only given delay values (namely 70, 80 and 90) are effective in our setting. We also test the derived bandwidth assured to a hypothetical critical core when interfering cores (e.g., generating a concurrent memory access workload) are present on the same machine. Our results can support designers by providing understanding of impact of the shared memory to enable predictable progress of safety critical applications in cloud environments., Comment: 8 pages, to appear in proceedings of "The 25th International Symposium On Real-Time Distributed Computing ISORC"

Virtualization of distributed real-time systems enables the consolidation of mixed-criticality functions on a shared hardware platform thus easing system integration. Time-triggered communication and computation can act as an enabler of safe hard real-time systems. A time-triggered hypervisor that activates virtual CPUs according to a global schedule can provide the means to allow for a resource efficient implementation of the time-triggered paradigm in virtualized distributed real-time systems. A prerequisite of time-triggered virtualization for hard real-time systems is providing access to a global time base to VMs as well as to the hypervisor. A global time base is the result of clock synchronization with an upper bound on the clock synchronization precision. We present a formalization of the notion of time in virtualized real-time systems. We use this formalization to propose a virtual clock condition that enables us to test the suitability of a virtual clock for the design of virtualized time-triggered real-time systems. We discuss and model how virtualization, in particular resource consolidation versus resource partitioning, degrades clock synchronization precision. Finally, we apply our insights to model the IEEE~802.1AS clock synchronization protocol and derive an upper bound on the clock synchronization precision of IEEE 802.1AS. We present our implementation of a dependent clock for ACRN that can be synchronized to a grandmaster clock. The results of our experiments illustrate that a type-1 hypervisor implementing a dependent clock yields native clock synchronization precision. Furthermore, we show that the upper bound derived from our model holds for a series of experiments featuring native as well as virtualized setups.

One of the primary sources of unpredictability in modern multi-core embedded systems is contention over shared memory resources, such as caches, interconnects, and DRAM. Despite significant achievements in the design and analysis of multi-core systems, there is a need for a theoretical framework that can be used to reason on the worst-case behavior of real-time workload when both processors and memory resources are subject to scheduling decisions. In this paper, we focus our attention on dynamic allocation of main memory bandwidth. In particular, we study how to determine the worst-case response time of tasks spanning through a sequence of time intervals, each with a different bandwidth-to-core assignment. We show that the response time computation can be reduced to a maximization problem over assignment of memory requests to different time intervals, and we provide an efficient way to solve such problem. As a case study, we then demonstrate how our proposed analysis can be used to improve the schedulability of Integrated Modular Avionics systems in the presence of memory-intensive workload., Comment: Accepted for publication in the IEEE Real-Time Systems Symposium (RTSS) 2018 conference