Internet of Things starts to integrate deeply our daily life through the comfort and the services that it offers. This technology made us already more intertwined with the external environment via deployed communicating devices. However, even though this proximity offers many advantages, it presents also strong security issues. The cyber-attack surface is significantly increased and the intruder impact is becoming more compromising. In this context, we present the preliminary results of our study toward the conception of a machine learning based solution against cyber-intrusions. The main purpose is to develop a resource-preserving solution with high precision of detection. We were interested in the IoTID20 dataset over which we experimented the most accurate machine learning models. The employed pre-processing approach that we propose pushes the machine learning models to the peak of their performances. Our study provides a comprehensive view of models efficiency with respect to detection rate, size and delay.

We propose the integration of blockchain with the OAuth standard. Two blockchains are considered for the implementation: Ethereum where ACL and token processing is performed on-chain and MultiChain where the processing is performed off-chain. The experiments show that the on-chain token generation provides an additional overhead compared to the off-chain one. However, it provides a secure token management and access traceability. The main advantage of our contribution is to propose a decentralized authorization system while relying on a standard specification.

Long Range Wireless Access Network (LoRaWAN) emerged as one of the promising Low Power Wide Area Networks (LPWAN) for IoT applications. It allows end-devices to reach a gateway and then the core network with a star topology in a wide area. Long Range (LoRa) transceivers send data packets according to a configuration or a set of parameter's values: Spreading Factor (SF), Payload size (PS), Bandwidth (BW) and Coding Rate (CR). These parameters must be fixed or adapted to application's requirements. Adaptive Data Rate (ADR) control system of LoRaWAN has been proposed to adapt modulation parameters dynamically based on the recent received packets. However, ADR control system doesn't adjust parameters considering the evolution of applications' Quality of Service (QoS) requirements. In this paper, we propose to cluster a set of LoRa transmission settings based on the measured QoS metrics such as Bit Error Rate (BER), Time on Air (ToA) and Received Signal Strength Indication (RSSI). We consider the set of settings' vectors as a cloud of points in a vector space while measured metrics are points' coordinates. Our method aims to map a set of LoRa transmission settings that offers the same QoS to the same cluster. We generate a set of transmission settings randomly and apply the Fuzzy C-Means (FCM) clustering algorithm on the resulting QoS metrics, Results show that the FCM clustering algorithm attribute membership values that best fit application requirements. This result could be used by LoRaWAN network servers to map each LoRa transmission setting to the application running on end devices.

In the context of smart grids, Demand Response (DR) is used to manage energy imbalance by smoothing consump¬tion peaks through voluntary rationing of energy by participants. However, it largely remains centralized and opaque with little to no traceability. To resolve this, we propose a blockchain-based framework in which a consortium of DR allotters and certify¬ing authorities maintain the blockchain. This brings in more transparency, traceability, and complete decentralization along with trustlessness, non-repudiation, and immutability. Further, the framework uses distinct components/concepts like Secure Elements, Escrow Accounts, Applied Smart Contracts in unison to fix the impediments of previous blockchain-based propositions. Next, we propose a fair and efficient DR allotment mechanism for a distributed DR marketplace whose execution time is less than 1 minute for more than 20,000 participants. Further, through simulations, we show the impact of different parameters on it and demonstrate its ability to delicately balance various paradigms of DR metrics. Finally, we conclude with our findings on system reliability and its inordinate effects on DR allotment metrics.

Coarse-grained programmable systolic architectures are designed to meet hard time constraints and provide high-performance computing. They consist of a set of programmable hardware resources with directed interconnections between them. The level of complexity of these architectures limits their re-usability. An automated mapping methodology is required to add a re-usability value to these architectures. In this work, we present a new list-scheduling based mapping methodology for coarse-grained programmable systolic architectures. We use a Directed Acyclic Graph to express the tasks and data dependency of the application as well as the hardware resources organization. We demonstrate that our approach can map different applications, provide a latency estimation and generate the configuration context. This approach could be the base for design space exploration and optimization tools for this family of architectures.

Bitcoin is a digital currency that has enticed considerable recognition as an alternative solution to the traditional payment system. Its success is mostly due to the use of Blockchain over a decentralised peer-to-peer network. Today's Bitcoin network has more than 10,000 online peers geographically scattered around the world. However, its topology is still unknown due to security concerns. In this paper, we design and implement a fast and efficient framework named BTCmap to discover and map the Bitcoin network topology. Our BTCmap framework includes two modules viz., a sniffer to collect the local addresses database of each peer, and a Bitcoin peer emulator to select neighbours and generate the topology. Our analyses show that to remain within 1% confidence interval, we have a window of 56 minutes 40 seconds for collecting the local address databases from all peers. Within this time duration, BTCmap requests more than 8200 reachable peers to map the real Bitcoin network topology. We present a Bitcoin topology snapshot captured on June 14, 2018, 15:53 CEST and show that this snapshot is connected.

In this paper, we consider the problem of guaranteeing real-time end-to-end transmission time for flows sent on a Network-on-Chip (NoC) with First-in First-out (FIFO) scheduling on each node. We show how to adapt the Trajectory approach, used in the context of Avionics Full DupleX switched Ethernet (AFDX) networks to characterize end-to-end transmission delays, to the context of NoC-based Systems-on-Chip (SoCs). We characterize the benefit of the Trajectory approach on an example.

We consider the problem of dimensioning real-time AFDX FIFO networks with a worst-case end-to-end delay analysis. The state-of-the-art has considered several approaches to compute these worst-case end-to-end delays. Among them, the Trajectory approach has received more attention as it has been shown to provide tight end-to-end delay upper bounds. Recently, it has been proved that current Trajectory analysis can be optimistic for some corner cases, leading in its current form, to certification issues. In this paper, we first characterize the source of optimism in the Trajectory approach on detailed examples. Then, we provide a correction to the identified problems. Two problems are solved: the first one is on the root cause of the underestimated time interval to compute delays of competing flows and a problem in the definition of the end-to-end delay computation. The second one is on the way that serialized frames are taken into account in the worst-case delay analysis.

Safety Critical systems are those systems which may endanger human lives if a failure occurs. This is the case, for example, in avionics and in modern vehicles. Security critical systems contains data or knowledge that should stay confidential. Depending on the impact of such revelation, the security level requested varies. In today's modern vehicles, many safety-critical subsystems are also security critical. These subsystems are usually interconnected through a network, each subsystem having different criticality levels. Problems come when low criticality (e.g. a multimedia application) subsystems need to exchange information with high criticality ones. If nothing is done, the low criticality subsystem inherits the requirements of the higher criticality system that it is communicating with, generating high certification costs. Such a problem increases in complex vehicles vetronic systems with advanced technologies such as United Nation peace protection armoured vehicles and some vehicles used by public services. This is not acceptable when certification is required for only a subset of subsystems. In this paper we define systems having heterogeneous communicating subsystems with multi-criticality and multi-security issues as systronic systems. This article contributes to an initial framework for the design of a systronic system with the concept of a connector that respects safety and security constraints. The goal of this connector is to control communications between different subsystems, while preserving the safety and the security of each subsystem without additional costs.

Current trends in the embedded systems field tend to collocate multiple functionalities upon a single computing platform, the aim being to reduce both the size and cost of embedded systems. Nevertheless, it is unlikely that all functionalities share the same level of criticality, and certification of the system has to be achieved using varying degrees of rigorousness. Typically, a task tau_i is guaranteed to meet its temporal constraints up to a criticality level that is equal to its own criticality. When those conditions are no longer met, i.e. when another higher priority task tau_j has its execution time that exceeds its Worst Case Execution Time (WCET) w.r.t. the criticality level of tau_i, a common approach is to suspend tau_i. However, in some cases, it may not be necessary to suspend tasks with a lower criticality immediately as they could still be executed without compromising the deadlines of high criticality tasks. As a step towards this aim, we propose a method, denoted Latest Completion Time (LCT), that allows lower criticality tasks to proceed with their execution as long as they do not prevent higher criticality tasks from meeting their deadlines. Furthermore, we show that tasks suspension can only be temporary, and prove that a particular definition of idle times can be used to reset the system's criticality level. Finally, we study the performances of our LCT mechanism w.r.t. the classical mechanism that suspends a task as soon as the system criticality level becomes higher than its own criticality.

We consider the problem of fixed priority scheduling of non-preemptive strict periodic tasks in conjunction with sporadic preemptive tasks. There are few studies about the scheduling problem combining these two kinds of tasks. Moreover, only few results are available on scheduling non-preemptive strict periodic tasks since their performance analysis gives low success ratios, except in the case of harmonic tasks. Also, strict periodic tasks are of great importance since they are in charge for example of sensors/actuators or feedback control functions which are all critical in feedback control systems. Such tasks must have the highest priorities in order to guarantee a correct behavior of the control system. Preemptive sporadic tasks can be used for non critical functions and have lower priorities. We first investigate the scheduling problem of non-preemptive strict periodic tasks by recalling an existing schedulability condition. This results in defining the first release times of strict periodic tasks that preserves the strict periodicity constraints. We show that the schedule of strict periodic tasks can have transient and permanent phases. Then, assuming that some non-preemptive strict periodic tasks have been scheduled, we characterize the release times of the sporadic tasks that maximize their worst case response times. We prove that these release times can be restricted to the permanent phase. For preemptive sporadic tasks, we extend the classical worst case response time computation to take into account non-preemptive strict periodic tasks. Finally, we consider the particular case where some of the sporadic tasks are alternate tasks to primary strict periodic tasks for fault-tolerance.

We focus on the real-time multiprocessor scheduling of periodic tasksets. We propose a new static priority scheduling algorithm based on the restricted-migration approach. Restricted-migration approach is a global scheduling approach for which the number of migrations is bounded just by one migration per job at most. Our algorithm uses the laxity of already admitted jobs to decide the admission of newly arrived jobs. We prove that this algorithm is predictable. We give a feasible interval and we propose a utilization bound for this algorithm. We also compare our algorithm to other global algorithms in terms of schedulability by simulations.

In this paper we study the partitioning approach for multiprocessor real-time scheduling. This approach seems to be the easiest since, once the partitioning of the task set has been done, the problem reduces to well understood uniprocessor issues. Meanwhile, there is no optimal and polynomial solution to partition tasks on processors. In this paper we analyze partitioning algorithms from several points of view such that for a given task set and specific constraints (processor number, task set type, etc.) we should be able to identify the best heuristic and the best schedulability test. We also analyze the influence of the heuristics on the performance of the uniprocessor tests and the impact of a specific task order on the schedulability. A study on performance difference between Fixed Priority schedulers and EDF in the case of partitioned scheduling is also considered.

In this paper, we propose a new approach for the partitioned Earliest Deadline First (EDF) scheduling of sporadic task systems. We consider the case of constrained task deadlines where the deadlines of the tasks are less than or equal to their periods. We introduce the concept of the EDF norm, for defining the space of WCET values that result in schedulable systems given fixed periods and relative deadlines. Based on this concept, it is possible to derive a necessary and sufficient feasibility condition to check whether EDF scheduling is valid for a given partitioning. The EDF norm has interesting convexity properties that permit using a Linear Programming approach to reduce the number of points at which the EDF norm needs to be checked. From the EDF norm, we derive a New Worst Fit Decreasing partitioning heuristic and compare its performance with two existing partitioning heuristics based on density partitioning and demand bound function approximation. We then compare the performance of the heuristic in terms of the resource augmentation paradigm.

We consider in this paper a partitioned real-time multiprocessor scheduling problem for sporadic task having constrained deadlines and a fixed priority scheduling. A major problem with partitioned scheduling is to determine if there exists a valid partitioning satisfying all the tasks deadlines. Many algorithms have tried to solve this problem. Most of them have considered first fit partitioning to minimize the number of processors required at the price of maximizing the load on each assigned processors. In this paper, we propose another partitioning strategy that aims at maximizing the allowance of WCETs of the tasks defined as the maximum extra duration that can be given to a task without compromising the task deadlines. The allowance-fit partitioning we propose uses the allowance criteria. We show that the partition found by our algorithm improves the temporal robustness of all the tasks of the system in the case of WCET overruns faults.

In this paper, we propose to extend real-time feasibility conditions taking into account kernel overheads for mixed preemptive periodic tasks with shared resources when tasks are scheduled with Fixed Priority FP/FIFO scheduling, where FIFO is used to arbitrate tasks having the same fixed priority. The kernel considered in this paper is an event driven OSEK kernel which uses the priority ceiling protocol to avoid deadlocks. The overheads are due to context switching, task activations and terminations, and task rescheduling. Periodic timers used to create the time base to implement periodic tasks and to the OSEK's services used to request and release resources also introduce kernel overheads. We compare the theoretical worst case response time obtained with kernel overheads to the response time obtained on a real event driven OSEK implementation.

In this paper, we consider a real-time application composed of graphs of subtasks. We present a mechanism for execution-overrun management in realtime systems. A subtask exceeding its prescribed execution duration can result in multiple failures (cascading effect). Yet, a task execution overrun does not necessarily lead to a deadline miss. Indeed, a system with enough free resources can self-stabilize without compromising the timeliness constraints of the subtasks in the system. We propose a mechanism based on the notion of allowance that enables to cope with execution overruns before a deadline miss. The task allowance depends on free resources. We show how to determine the maximum allowance duration a subtask can be allowed to proceed with it execution without compromising the real-time constraints of all the subtasks. We show how to determine the maximum allowance of a subtask from the feasibility conditions of sporadic graphs of subtasks scheduled with a preemptive fixed priority scheduling

In this paper we show how to take into account kernel overheads in classical real-time feasibility conditions for fixed priority (FP) scheduling where tasks having the same fixed priority are scheduled FP/FIFO. We consider the periodic task model with arbitrary deadlines and an event driven OSEK kernel. The feasibility conditions are based on the worst case response time computation of the tasks. We identify the sources of kernel overheads that influence the response time of the tasks. In such a system the overheads are due to the context switching that activates/terminates and reschedules tasks and to the granularity of the periodic timer used to implement the periodic task model. We show how to take into account those overheads in the classical FP/FIFO feasibility conditions. We compare the worst case response time obtained with kernel overhead to the response time obtained on a real event driven OSEK implementation. We show that the kernel overheads cannot be neglected and that the theoretical results are valid and can be used for a real-time dimensioning.

This paper considers the problem of preemptive hard real-time scheduling of sporadic graph tasks on a uniprocessor with Earliest Deadline First (EDF). A sporadic graph task associated to an application is characterized by a set of subtasks under precedence constraints with deadline constraints. We study feasibility conditions based on the computation of the worst case response time of any subtask of a graph task. We investigate the late deadline constraint.

More and more applications supported by MANETs require quality of service (QoS). QoS can be expressed in terms of bandwidth or delay. In this paper, we study the impact of the scheduling policy used by each MANET node on the QoS granted to flows. We first show how to use WCBQ, a class based queueing scheduling, in a wireless network. We then compare the performances achieved by WCBQ and priority queueing. Simulation results show that WCBQ protects flows with low throughput. All flows receive a bandwidth proportionally to their weights. WCBQ tends to minimize the variations of the average bandwidth granted to a flow as well as the end-to-end delay. We finally propose two solutions allowing the coexistence of flows with delay requirements and bandwidth requirements. These solutions guarantee a low delay to low throughput flows without penalizing the bandwidth granted to other flows

This paper considers the problem of hard real-time fixed priority scheduling of sporadic graph tasks on a uniprocessor. We study feasibility conditions based on the computation of the worst case response times of the graph tasks. A sporadic graph task associated to an application is characterized by a set of subtasks under precedence constraints. We investigate the late deadline constraint. We extend already published papers established in the preemptive case to the nonpreemptive case and to the hybrid case of an application composed of preemptive and nonpreemptive subtasks.

In this paper, we focus on a Bluetooth piconet, analyzing its ability to support the quality of service (QoS) requirements defined by QoS sensitive applications. We consider two QoS parameters: (i) an application constraint denoting the importance degree of a message, and (ii) an end-to-end delivery deadline. The QoS perceived by the application depends on the scheduling algorithms chosen at the medium access layer. We determine the average response time of a message obtained with three different scheduling algorithms. This evaluation takes into account message segmentation in the Bluetooth MAC layer. The first scheduling is the Bluetooth native scheduling called One-Round Robin (1-RR). The second is a variant of 1-RR where the local scheduling is based on the two QoS parameters considered. The last one, called Class-Based Earliest Deadline First (CB-EDF), takes into account the two QoS parameters both at the local level and at the global level.

We investigate the problem of guaranteed end-to-end response times for video/audio broadcast systems. A flow is sent by a video/audio source and recorded by video servers for local clients. Because of transmission delay variability, the video/audio flows suffer release jitter. This end-to-end guarantee is based on an admission control run only by the nodes belonging to the path of the requesting flow. The requesting flow is accepted only if it does not endanger the guarantees already granted. All nodes use a non-preemptive non-idling scheduling: video/audio servers use earliest deadline first (EDF) and other nodes use fixed priority/highest priority first (FP/HPF).

A solution for a hard real time scheduling problem in a distributed system is designed and proved. The constraints of our problem are first to preserve consistency even in the presence of concurrency and second to preserve the order of task releases, provided that task release times differ more than clock precision. That is achieved by FIFO based scheduling. The feasibility conditions resulting from the worst case response time analysis of each task set are given. The solution complexity is shown to be pseudo polynomial.