In multiprocessor-based real-time systems, main memory is identified as a major bottleneck in the worst-case timing analysis of tasks. Phased execution models such as the 3-phase task model, i.e., that divides the execution of tasks into distinct computation and memory phases, have shown to be a good candidate to tackle the memory contention problem. The 3-phase execution model in particular has gained much attention from both academia and industry as it limits when tasks can access main memory to pre-defined phases. Information on when those phases may happen and their length can then be leveraged to build a fine-grained memory contention analysis. However, the existing work that focus on the memory contention analysis for 3-phase tasks may overestimate the memory contention caused by interfering write requests. This yields pessimistic bounds on the total memory contention suffered by tasks which in turn leads to pessimistic worst-case execution time (WCET) and worst-case response time (WCRT) bounds. In this work, we improve the state-of-the-art memory contention analysis for 3-phase tasks by (i) tightly bounding the memory contention that can be suffered due to write requests; and (ii) providing a new memory contention-aware WCET analysis.

The 3-phase task execution model has shown to be a good candidate to tackle the memory bus contention problem. It divides the execution of tasks into computation and memory phases that enable a fine-grained memory bus contention analysis. However, existing works that focus on the bus contention analysis for 3-phase tasks, neglect the fact that memory bus contention strongly relates to the number of bus/memory requests generated by tasks, which, in turn, depends on the content of the cache memories during the execution of those tasks. These existing works assume that the worst-case number of bus/memory requests will be generated during all the memory phases of all tasks, irrespective of the already existing content in the cache memory. This overestimates the memory bus contention of tasks, leading to pessimistic worst-case response time (WCRT) bounds. This work proposes a holistic approach towards bus contention analysis for 3-phase tasks by (1) deriving an upper bound on the actual cache misses of tasks that lead to bus/memory requests; (2) improving State-of-the-Art (SOTA) bus contention analysis of two bus arbitration schemes that dominate all existing works on the bus contention analysis for 3-phase tasks; and (3) performing an extensive experimental evaluation under different settings to compare the proposed analysis against the SOTA. Results show that incorporating a tighter bound on the number of cache misses of tasks into the bus contention analysis can lead to a significant improvement in the task set schedulability.

The sharing of main memory among concurrently executing tasks on a multicore platform results in increasing the execution times of those tasks in a non-deterministic manner. The use of phased execution models that divide the execution of tasks into distinct execution and memory phase(s), e.g., the PRedictable Execution Model (PREM) and the 3-Phase task model, along with Memory Centric Scheduling (MCS) present a promising solution to reduce main memory interference among tasks.Existing works in the state-of-the-art that focus on MCS have considered (i) a TDMA-based memory scheduler, i.e., tasks’ memory requests are served under a static TDMA schedule, and (ii) Processor-Priority (PP) based memory scheduler, i.e., tasks’ memory requests are served depending on the priority of the processor/core on which the task is executing. This paper extends MCS by considering a Task-Priority (TP) based memory scheduler, i.e., tasks’ memory requests are served under a global priority order depending on the priority of the task that issues the requests. We present an analysis to bound the total memory interference that can be suffered by the tasks under the TPbased MCS. In contrast to the recent works on MCS that considers non-preemptive tasks, our analysis considers limited preemptive scheduling. Additionally, we investigate the impact of different preemption points on the memory interference of tasks. Experimental results show that our proposed TP-based MCS can significantly reduce the memory interference that can be suffered by the tasks in comparison to the PP-based MCS.

Commercial-off-the-shelf (COTS) multicore processors have become a preferable choice for modern systems to meet the increasing functionalities and computational demand of modern applications. However, the adoption of multicore platforms in hard real-time systems, i.e., systems that run applications with stringent timing requirements, is still under scrutiny. The main challenge that hinders the use of COTS multicore platforms in hard real-time systems is their unpredictability, which originates from the sharing of different hardware resources. A task executing on one core of a multicore platform has to compete with other co-running tasks (running on other cores) to access hardware resources such as the last-level cache (LLC), the interconnect (e.g., memory bus), and the main memory. This competition leads to inter-core contention which can significantly impact the Worst-Case Execution Time (WCET) and Worst-Case Response Time (WCRT) of tasks.

Multicore platforms are being increasingly adopted in Cyber-Physical Systems (CPS) due to their advantages over single-core processors, such as raw computing power and energy efficiency. Typically, multicore platforms use a shared memory bus that connects the cores to the off-chip main memory. This sharing of memory bus may cause tasks running on different cores to compete for access to the main memory whenever data/instructions are need to be read/written from/to the main memory. Such competition is problematic, as it may cause variations in the execution time of tasks in a non-deterministic way. To reduce the complexity of analyzing this problem, the 3-phase task model was proposed that divides tasks’ executions into distinct memory and execution phases. The distinctive memory phases are then scheduled to eliminate/minimize main memory contention between concurrently executing tasks. However, 3-phase tasks running on different cores may still compete to access the shared memory bus/main memory in order to execute memory phases. This paper presents a partitioned scheduling-based approach that allows one to derive memory bus contention-aware worst-case response time of tasks that follow the 3-phase task model. In particular, the bus-contention analysis is derived by considering two memory access models, i.e., (i) dedicated memory access model, where a core having allowed to access the main memory via memory bus is permitted to execute more than one memory phase, and (ii) fair memory access model, that restrict each core to execute only one memory phase in its allocated bus access. Both these models represent different system and application requirements, and the resulting bus contention of tasks may vary depending on the considered model. To evaluate the effectiveness of the proposed bus contention analysis, we compare its performance against an existing analysis in the state-of-the-art by performing (i) case-study experiments, using benchmarks from the Mälardalen Benchmark suite, and (ii) empirical evaluation using synthetic task sets. Results show that our proposed analysis can improve task set schedulability of 3-phase tasks by up to 88 percentage points.

Multicore platforms are being increasingly adopted in Cyber-Physical Systems (CPS) due to their advantages over single-core processors, such as raw computing power and energy efficiency. Typically, multicore platforms use a shared system bus that connects the cores to the memory hierarchy (including caches and main memory). However, such hierarchy causes tasks running on different cores to compete for access to the shared system bus whenever data reads or writes need to be made. Such competition is problematic as it may cause large variations in the execution time of tasks in a non-deterministic way. This paper presents an analysis that allows one to derive bus contention-aware worst-case response-time of tasks that follow the 3-phase task model executing under partitioned scheduling.

Developing complex safe and secure Cyber-Physical Systems (CPS) applications for the automotive domain is typically a complex task, due to the criticality inherent to this domain. Considering such known complexity of the development process, we propose a novel solution that aims to provide a quasi-automatic' integration process between the different components of such CPS systems via the support of a Domain Specific Language (DSL) that provides several views of the system, abstracting away the more technical implementation details, while imposing system properties and restrictions that have the potential to be formally verified (either statically or at run-time) during design, and facilitates the process of customization and quasi-automatic build and deployment processes. In this paper, we briefly analyze the tools that are available and that cover partially the characteristics of our envisioned DSL, describe its building blocks, and show how it can be applied in a small, yet sufficiently complex CPS application whose architecture is very close to what we may expect for the modern and future generation of CPS application in the automotive domain.

The Event Handler - a publish-subscribe broker implemented over REST/HTTP(S) - is an auxiliary system of the Arrowhead framework for IoT applications. During this work we found that the existing implementation of the Event Handler suffers from serious performance issues. This paper describes the reengineering effort that ultimately enabled it to reach much more acceptable levels of performance, by using appropriate software configurations and design patterns. Additionally, we also illustrate how this enhanced version of the Event Handler can be modeled using Petri nets, to depict the performance impact of different thread pool configurations and CPU core availability. The main objective of this modeling process is to enable the estimation of the system's performance to guarantee the required quality of service.

The Arrowhead Framework, a SOA-based framework for IoT applications, provides the Event Handler system: a publish/subscribe broker implemented with REST/HTTP(S). However, the existing implementation of the Event Handler suffers from message latency problems that are not acceptable for industrial applications. Thus, this paper describes the refactoring process of this system that enabled it to reach acceptable levels of latency.

Scheduling real-time applications on general purpose multicore platforms is a challenging problem from a timing analysis perspective. Such platforms expose uncontrolled sources of interference whenever concurrent accesses to memory are performed. The non-deterministic bus and memory access behavior complicates the estimations of applications' worst-case execution times (WCET). The 3-phase task model seems a good candidate to circumvent the uncontrolled sources of interference by isolating concurrent memory accesses. A task is divided in three successive phases; first, the task loads its instruction and data in a local memory, then it executes non-preemptively using those pre-loaded instructions and data, and finally, the modified data are pushed back to main memory. Following this execution model, tasks never access the bus during their execution phase. Instead, all the bus accesses are performed during the first and third phases. In this paper, we focus on the global fixed-priority scheduling of the 3-phase task model. A new schedulability test is derived by modelling the interference happening on the bus rather than the interference on the cores as in the state-of-the-art techniques. The effectiveness of the test is evaluated by comparing it against the state-of-the-art.

Commercial-of-the-shelf based multi-core systems present timing anomalies that cannot be ignored by the real-time systems community due to their unpredictable behaviour. These timing anomalies, often caused by applications' uncontrolled accesses to shared resources such as the components in the memory hierarchy or in the I/O subsystem, introduce interference that may lead to deadline misses if the problem is neglected. The Acquisition Execution Restitution (AER) execution model was previously proposed to circumvent this problem and, therefore, mitigate inter-task interference. In this model, applications decouple communication (acquisition and restitution phases) from the actual execution in a way that at most one acquisition or restitution phase is in execution at any instant of time while the execution phase of different tasks can progress in parallel on multiple cores. Thus, keeping each task's derived worst-case execution time closer to the one measured in isolation. In this paper, we study the AER execution model and compare it against a global Earliest Deadline First (EDF) approach where interferences are considered. Our results show that a priority assignment heuristic which assigns the priorities based on the tasks' periods dominates all the other proposed heuristics and that due to interference it can also schedule task sets which are not schedulable by using the global EDF approach.

This paper explores the behavior of parallel fork-join tasks on multicore platforms by resorting to a semi-partitioned scheduling model. This model offers a promising framework to embedded systems which are subject to stringent timing constraints as it provides these systems with very interesting properties. The proposed approach consists of two stages -- an offline stage and an online stage. During the offline stage, a multi-frame task model is adopted to perform the fork-join task-to-core mapping so as to improve the schedulability and the performance of the system, and during the online stage, work-stealing is exploited among cores to improve the system responsiveness as well as to balance the execution workload. The objective of this work is twofold: (1) to provide an alternative technique that takes advantage of the semi-partitioned scheduling properties by offering the possibility to accommodate fork-join tasks that cannot be scheduled in any pure partitioned environment, and (2) to reduce the migration overhead which has shown to be a traditional major source of non-determinism in global approaches. The simulation results show an improvement of the proposed approach over the state-of-the-art of up to 15% of the average response-time per task set.