REACTION 2013. Second 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. Second International Workshop on Real-time and distributed computing in emerging applications

Table of contents

Preliminary pages

Message from the General Chairs

Session A - Middelware and distributed systems

Session B - Scheduling in distributed systems and multiprocessors

Session C - Cloud computing


Recent Submissions

Now showing 1 - 9 of 9
  • Publication
    Message from the General Chairs
    (Universidad Carlos III de Madrid, 2013-12) García Valls, María Soledad; Cucinotta, Tommaso; García Valls, Marisol; Cucinotta, Tommaso
  • Publication
    Preliminary pages [REACTION 2013, co-located with IEEE RTSS]
    (Universidad Carlos III de Madrid, 2013-12) García Valls, Marisol; Cucinotta, Tommaso
  • Publication
    A Feasible Configuration of AFDX Networks for Real-Time Flows in Avionics Systems
    (Universidad Carlos III de Madrid, 2013-12-03) An, Dongha; Jeon, Hyun Wook; Kim, Kyong Hoon; Kim, Ki-Il; García Valls, Marisol; Cucinotta, Tommaso
    AFDX (Avionics Full Duplex Switched Ethernet) Networks have been proposed to meet unique ADN (Aircraft Data Networks) characteristics and then standardized as a Part 7 in ARNIC 664. As for this new communication technology, some research works have been conducted to address design issues such as optimizing virtual links as well as analytic modeling including response time. Despite of their research efforts, configuration problem for both MTU (Maximum Transmission Unit) and BAG (Bandwidth Allocation Gap) over virtual links in AFDX networks remains unsolved yet. In this paper, we propose how to set MTU and BAG value on each virtual link according to both application requirements and AFDX switch constraints. We define a new problem of feasible configurations of virtual links in an AFDX switch and propose an algorithm to derive feasible BAG and MTU pairs based on the branch-and-bound technique. Throughout simulations, we evaluate the proposed algorithm and analyze the effect of parameters in AFDX networks.
  • Publication
    Experimental Evaluation of the Real-Time Performance of Publish-Subscribe Middlewares
    (Universidad Carlos III de Madrid, 2013-12-03) Rizano, Tizar; Abeni, Luca; Palopoli, Luigi; García Valls, Marisol; Cucinotta, Tommaso
    The integration of the complex network of modules composing a modern distributed embedded systems calls for a middleware solution striking a good tradeoff between conflicting needs such as: modularity, architecture independence, re-use, easy access to the limited hardware resources and ability to respect real–time constraints. Several middleware architectures proposed in the last years offer reliable and easy to use abstractions and intuitive publish-subscribe mechanism that can simplify system development to a good degree. However, a complete compliance with the different requirements of assistive robotics application (first and foremost real–time constraints) remains to be investigated. This paper evaluates the performance of these solutions in terms of latency and scalability.
  • Publication
    Towards the integration of data-centric distribution technology into partitioned embedded systems
    (Universidad Carlos III de Madrid, 2013-12-03) Pérez, Héctor; Gutiérrez, J. Javier; García Valls, Marisol; Cucinotta, Tommaso
    This work proposes an architecture to enable the use of data-centric real-time distribution middleware in partitioned embedded systems based on a hypervisor. Partitioning is a technique that provides strong temporal and spatial isolation, thus allowing mixed-criticality applications to be executed in the same hardware. The proposed architecture not only enables transparent communication among partitions, but it also facilitates the interconnection between partitioned and nonpartitioned systems through distribution middleware. Preliminary results show that hypervisor technology provides low overhead and a reasonable trade-off between temporal isolation and performance.
  • Publication
    Resource Management for Service Level Aware Cloud Applications
    (Universidad Carlos III de Madrid, 2013-12-03) Klein, Cristian; Hernández-Rodríguez, Francisco; Maggio, Martina; Arzen, Karl-Erik; García Valls, Marisol; Cucinotta, Tommaso
    Resource allocation in clouds is mostly done assuming hard requirements, time-sensitive applications either receive the requested resources or fail, Given the dynamic nature of workloads, guaranteeing on-demand allocations requires large spare capacity. Hence, one cannot have a system that is both reliable and efficient. To mitigate this issue, we introduce service-level awareness in clouds, assuming applications contain some optional code that can be dynamically deactivated as needed. We propose a resource manager that allocates resources to multiple service-level-aware applications in a fair manner. To show the practical applicability, we implemented service-level-aware versions of RUBiS and RUBBoS, two popular cloud benchmarks, together with our resource manager. Experiments show that service-level awareness helps in withstanding flash-crowds or failures, opening up more flexibility in cloud resource management.
  • Publication
    Benchmarking communication middleware for cloud computing virtualizers
    (Universidad Carlos III de Madrid, 2013-12-03) García Valls, María Soledad; Basanta Val, Pablo; Serrano Torres, Rosbel; García Valls, Marisol; Cucinotta, Tommaso
    Virtualization technologies typically introduce additional overhead that is specially challenging for specific domains such as real-time systems. One of the sources of overhead are the additional software layers that provide parallel execution environments which reduce the effective performance given by the infrastructure. This work identifies the factors to be analysed by a benchmark for performance evaluation of a virtualized middleware. It provides the set of benchmark tests that evaluate empirically the overhead and stability on a trendy communication middleware, DDS (Data Distribution System for Real-Time), which enables message transmissions via publisher-subscriber (P/S) interactions. Two different implementations, RTI and OpenSplice, have been analysed over a general purpose virtual machine monitor to evaluate their behavior on a client-server application. Obtained results have provided initial execution clues on the performance that a virtualized communication middleware like DDS can exhibit.
  • Publication
    Run-time Support for Real-Time Multimedia in the Cloud
    (Universidad Carlos III de Madrid, 2013-12-03) Cucinotta, Tommaso; Oberle, Karsten; Stein, Manuel; Domschitz, Peter; Mullender, Sape; García Valls, Marisol; Cucinotta, Tommaso
    This paper summarizes key research findings in the area of real-time performance and predictabil- ity of multimedia applications in cloud infrastruc- tures, namely: outcomes of the IRMOS European Project, addressing predictability of standard vir- tualized infrastructures; Osprey, an Operating Sys- tem with a novel design suitable for a multitude of heterogeneous workloads including real-time soft- ware; MediaCloud, a novel run-time architecture for offering on-demand multimedia processing facil- ities with unprecedented dynamism and flexibility in resource management. The paper highlights key research challenges ad- dressed by these projects and shortly presents ad- ditional questions lying ahead in this area.
  • Publication
    Task Partitioning and Priority Assigment for Hard Real-Time Distributed System
    (Universidad Carlos III de Madrid, 2013-12-03) Garibay-Martínez, Ricardo; Nelissen, Geoffrey; Ferreira, Luis Lino; Luis Miguel, Pinho; García Valls, Marisol; Cucinotta, Tommaso
    The partitioning of flxed-priority hard real-time tasks and messages in a distributed system is a well know NP-hard problem. Therefore, there are no methods that provide an optimal solution in polynomial time. In this paper, we propose the Distributed using Optimal Priority Assignment (DOPA) heuristic, which simultaneously solves the problem of assigning task to processors and assigning priorities to tasks. DOPA makes use of Audsley's ptimal Priority Assignment (OPA) algorithm to assign priorities to tasks and messages. However, in order to use the OPA algorithm for task sets with dependencies, we first transform the task set into a set of independent tasks by imposing intermedíate deadlines. The experimental results show how the utilisation of the OPA algorithm increases in average the number of schedulable tasks and messages in a distributed system when compared to the utilisation of the Deadline Monotonic (DM) priority assignment usually used in other works.