REACTION 2014. Third IEEE International Workshop on Real-time and distributed computing in emerging applications

Permanent URI for this collection

Archivo Abierto Institucional de la Universidad Carlos III de Madrid. 3rd IEEE International Workshop on Real-time and distributed computing in emerging applications

Table of contents

Preliminary pages

Message from the General Chairs

Cloud computing

CPS, network scheduling, and middleware technology

Real-time scheduling and multicore


Recent Submissions

Now showing 1 - 10 of 10
  • Publication
    Time Triggered Scheduling Analysis for Real-Time Applications on Multicore Platforms
    (Universidad Carlos III de Madrid, 2014-11) Freier, Matthias; Chen, Jian-Jia; Cucinotta, Tommaso; Pautet, Laurent; García Valls, Marisol
    Scheduling of real-time applications for multicore platforms has become an important research topic. For analyzing the timing satisfactions of real-time tasks, most researches in the literature assume independent tasks. However, industrial applications are usually with fully tangled dependencies among the tasks. Independence of the tasks provides a very nice abstraction, whereas dependent structures due to the tangled executions of the tasks are closer to the real systems. This paper studies the scheduling policies and the schedulabil-ity analysis based on independent tasks by hiding the execution dependencies with additional timing parameters. Our scheduling policy relates to the well-known periodic task model, but in contrast, tasks are able to communicate with each other. A feasible task set requires an analysis for each core and the communication infrastructure, which can be performed indi-vidually by decoupling computation from communication in a distributed system. By using a Time-Triggered Constant Phase (TTCP) scheduler, each task receives certain time-slots in the hyper-period of the task set, which ensures a time-predictable communication impact. In this paper, we provide several algorithms to derive the time-slot for each task. Further, we found a fast heuristic algorithm to calculate the time-slot for each task, which is capable to reach a core utilization of 90% by considering typical industrial applications. Finally, experiments show the effectiveness of our heuristic and the performance in different settings.
  • Publication
    GEN4MAST: A Tool for the Evaluation of Real-Time Techniques Using a Supercomputer
    (Universidad Carlos III de Madrid, 2014-11) Rivas, Juan M.; Gutiérrez, J. Javier; González Harbour, Michael; Cucinotta, Tommaso; Pautet, Laurent; García Valls, Marisol
    The constant development of new approaches in real-time systems makes it necessary to create tools or methods to perform their evaluations in an efficient way. It is not uncommon for these evaluations to be constrained by the processing power of current personal computers. Thus, it is still a challenging issue to know whether a specific technique could perform better than another one, or the improvement remains invariable in all circumstances. In this paper we present the GEN4MAST tool, which can take advantage of the performance of a supercomputer to execute longer evaluations that wouldn’t be possible in a common computer. GEN4MAST is built around the widely used MAST tool, automating the whole process of distributed systems generation, execution of the requested analysis or optimization techniques, and the processing of the results. GEN4MAST integrates several generation methods to create realistic workloads. We show that the different methods can have a great impact on the results of distributed systems.
  • Publication
    Data-centric distribution technology in ARINC-653 systems
    (Universidad Carlos III de Madrid, 2014-11) Pérez, Héctor; Gutiérrez, Javier; Cucinotta, Tommaso; Pautet, Laurent; García Valls, Marisol
    Standard distribution middleware has recently emerged as a potential solution to interconnect distributed systems in the avionics domain, as it would bring important benefits throughout the software development process. A remaining challenge, however, is reducing the complexity associated with current distribution standards which leads to prohibitive certification costs. To overcome this complexity, this work explores the use of the DDS distribution standard on top of a software platform based on the ARINC-653 specification. Furthermore, it discusses how both technologies can be integrated in order to apply them in mission and safety-critical scenarios.
  • Publication
    Heuristic Algorithm for Virtual Link Configuration in AFDX Networks
    (Universidad Carlos III de Madrid, 2014-11) Cha, YoungJun; Kim, Ki-Il; Cucinotta, Tommaso; Pautet, Laurent; García Valls, Marisol
    As the AFDX networks have been increasingly employed for airborne networks, much research works have been conducted to support real-time service in a deterministic way. However, since they assumed the preconfigured networks where all involved parameters were already determined, the impact of configuration algorithm is not well explored. To solve this problem, in this paper, we focus on how to reduce the required bandwidth by configuring virtual link which logically consists of at least one or more application flows. To achieve this, new heuristic algorithms have been proposed by applying well-known greedy approach while taking essential constraints of AFDX networks into account. To evaulate the performance of proposed scheme, diverse case studies for airborne application flows are concerned and their number of virtual links as well as required bandwidth are compared.
  • Publication
    Incremental Latency Analysis of Heterogeneous Cyber-Physical Systems
    (Universidad Carlos III de Madrid, 2014-11) Delange, Julien; Feiler, Peter; Cucinotta, Tommaso; Pautet, Laurent; García Valls, Marisol
    Cyber-Physical Systems, as used in automotive, avionics, or aerospace domains, have critical real-time require-ments. Time-related issues might have important impacts and, as these systems are becoming extremely software-reliant, validate and enforcing timing constraints is becoming difficult. Current techniques are mainly focused on validating these constraints late by using integration tests and tracing the system execution. Such methods are time-consuming and labor-intensive and, discovering timing issue late in the development process might incur significant rework efforts. In this paper, we propose an incremental model-based ap-proach to analyze and validate timing requirements of cyber-physical systems. We first capture the system functions, its related latency requirements and validate the end-to-end latency at a high level. This functional architecture is then refined into an implementation deployed on an execution platform. As system description is evolving, the latency analysis is being refined with more precise values. Such an approach provide latency analysis from a high level specification without having to implement the system, saving potential re-engineering efforts. It also helps engineers to select appropriate execution platform components or change the deployment strategy of system functions to ensure that latency requirements will be met when implementing the system.
  • Publication
    Towards a Performance Interference-aware Virtual Machine Placement Strategy for Supporting Soft Real-time Applications in the Cloud
    (Universidad Carlos III de Madrid, 2011-11) Caglar, Faruk; Shekhar, Shashank; Gokhale, Aniruddha; Cucinotta, Tommaso; Pautet, Laurent; García Valls, Marisol
    It is standard practice for cloud service providers (CSPs) to overbook physical system resources to maximize the resource utilization and make their business model more profitable. Resource overbooking can lead to performance interference, however, among the virtual machines (VMs) hosted on the physical resources causing performance un-predictability for soft real-time applications hosted in the VMs, which is unacceptable to these applications. Balancing these conflicting requirements needs a careful design of the placement strategies for hosting soft real-time applications such that the performance interference effects are minimized while still allowing resource overbooking. These placement decisions cannot be made offline because workloads change at run time. Moreover, satisfying the priorities of collocated VMs may require VM migrations, which require an online solution. This paper presents a machine learning-based, online placement solution to this problem where the system is trained using a publicly available trace of a large data center owned by Google. Our approach first classifies the VMs based on their historic mean CPU and memory usage, and performance features. Subsequently, it learns the best patterns of collocating the classified VMs by employing machine learning techniques. These extracted patterns are those that provide the lowest performance interference level on the specified host machines making them amenable to hosting soft real-time applications while still allowing resource overbooking.
  • Publication
    Real-Time Big Data: the JUNIPER Approach
    (Universidad Carlos III de Madrid, 2014-11) Audsley, N. C.; Chan, Y.; Gray, I.; Wellings, A. J.; Cucinotta, Tommaso; Pautet, Laurent; García Valls, Marisol
    Cloud computing offers the possibility for Cyber-Physical Systems (CPS) to offload computation and utilise large stored data sets in order to increase the overall system utility. However, for cloud platforms and applications to be effective for CPS, they need to exhibit real-time behaviour so that some level of performance can be guaranteed to the CPS. This paper considers the infrastructure developed by the EU JUNIPER project for enabling real-time big data systems to be built so that appropriate guarantees can be given to the CPS components. The technologies developed include a real-time Java programming approach, hardware acceleration to provide performance, and operating system resource manage-ment (time and disk) based upon resource reservation in order to enhance timeliness.
  • Publication
    A Multilevel I/O Tracer for Timing and Performance Analysis of Storage Systems in IaaS Cloud
    (Universidad Carlos III de Madrid, 2011-11) Ouarnough, Hamza; Boukhobza, Jalil; Singhoff, Frank; Cucinotta, Tommaso; Pautet, Laurent; García Valls, Marisol
    Data centers are more and more relying on hybrid storage systems consisting of flash memory based storage devices and traditional hard disk drives. Optimal data placement in such hybrid storage systems is a very important issue in the domain of cloud computing and virtualization. This is specially the case when users need that storage systems enforce Quality of Service requirements on I/Os performed, for example for multimedia applications. To characterize Virtual Machine (VM) I/O workload properties such as timing predictability or throughput, monitor-ing services are necessary on such new architectures. This article presents a multilevel I/O tracer for virtual machines that relies on and complement different state-of-the-art tools. It produces I/O traces at different levels of the Linux I/O software stack. The I/O tracer gives an exhaustive information that allows administrators to precisely characterize virtual machine I/O behavior in terms of percentage of read/write I/Os, percentage of random/sequential, I/O request inter-arrival time, etc. This tool is the first piece towards a middleware whose purpose is to meet user QoS requirements thanks to optimal data placement and migration policies in a hybrid storage system in the context of an IaaS Cloud.
  • Publication
    Message from the General Chairs
    (Universidad Carlos III de Madrid, 2014-11) García Valls, María Soledad; Cucinotta, Tommaso; Pautet, Laurent; Cucinotta, Tommaso; Pautet, Laurent; García Valls, Marisol
  • Publication
    Preliminary pages
    (Universidad Carlos III de Madrid, 2014-11) García Valls, Marisol; Cucinotta, Tommaso; Pautet, Laurent