Section: New Results

Quality-of-Service

Self-adaptive MAC protocol for both QoS and energy efficiency

Participants : Kévin Roussel, Shuguo Zhuo, Olivier Zendra, Ye-Qiong Song [contact] .

Three main contributions have been made this year. Firstly iQueue-MAC has been extended to work on both single channel mode and multi-channel mode, improving its throughput performance [11] . Secondly, S-CoSenS and iQueue-MAC our previously designed protocols have been implemented on RIOT OS over MSP430-based motes. Our contribution consists in developing a port of RIOT OS on the MSP430 micro-controller and demonstrating that RIOT OS offers rich and advanced real-time features, especially the simultaneous use of as many hardware timers as the underlying platform (micro-controller), which are fundamental features to implement high performance MAC protocols [41] . The Cooja/MSPSim network simulation framework is widely used for developing and debugging, but also for performance evaluation of WSN projects. Our third contribution shows that Cooja is not limited only to the simulation of the Contiki OS based systems and networks, but can also be extended to perform simulation experiments of other OS based platforms, especially that with RIOT OS. Moreover, when performing our own simulations with Cooja and MSPSim, we observed timing inconsistencies with identical experimentations made on actual hardware. Such inaccuracies clearly impair the use of the Cooja/MSPSim framework as a performance evaluation tool, at least for time-related performance parameters. The detailed results of our investigations on the inaccuracy problems, as well as the consequences of this issue, and possible ways to fix or avoid it are available in [42] . Part of this work has been supported by PIA LAR project.

End-to-end delay modeling and evaluation in wireless sensor networks

Participants : François Despaux, Abdelkader Lahmadi, Ye-Qiong Song [contact] .

Probabilistic end-to-end performance guarantee may be required when dealing with real-time applications. As part of ANR QUASIMODO project, we are dealing with Markov modeling of multi-hop networks running duty-cycled MAC protocols. One of the problems of the existing Markovian models resides in their strong assumptions that may not be directly used to assess the end-to-end delay in practice. In particular, realistic radio channel, capture effect and OS-related implementation factors are not taken into account. We proposed to explore a new approach combining code instrumentation and Markov chain analysis. In [32] we propose a novel approach to obtain the Markov chain model of sensor nodes by means of Process Mining techniques through the analysis of MAC protocol execution traces for a given traffic scenario. End to end delay is then computed based on this Markov chain. Experimentations were done using IoT-LAB testbed platform. Comparisons in terms of delay have been presented for two different metrics of the RPL protocol (hop count and ETX). The overall approach and its generalization using non-linear regression techniques in terms of traffic rate are detailed in the PhD thesis of François Despaux defended in September 2015 [1] .

Dynamic resource allocation in network virtualization

Participants : Mohamed Said Seddiki, Mounir Frikha [SupCom, Tunis, Tunisie] , Ye-Qiong Song [contact] .

This work has been carried out as part of a co-supervised PhD thesis between University of Lorraine and SupCom Tunis.

The objective of this research topic is to develop different resource allocation mechanisms in Network Virtualization, for increasing the QoS guarantee. Firstly, we demonstrated the potential of SDN in the QoS management of a virtualized home network (VN). We proposed and implemented ”FlowQoS”, a mechanism that can be deployed by an Internet Service Provider in the last-mile hop or in the home gateway. Performance measurements show that this solution can share bandwidth between applications according to user-defined configuration to guarantee QoS for each active traffic. The second contribution is the modeling of the interaction between service providers and infrastructure providers using game theory framework to offer dynamic sharing of physical infrastructure across multiple VN with different QoS requirements. We presented a set of non-cooperative games to model the negotiation phase and the dynamic allocation of nodes and physical links for each deployed VN[10] . Finally we proposed a predictive approach that allows an adaptive control of bandwidth allocation in order to reduce the packet delays for a given VN on each physical link. The last two contributions offer dynamic sharing models of physical infrastructure resources while guaranteeing the QoS for each VN.

The overall approach is detailed in the PhD thesis of Said Seddiki defended in April 2015 [2] .

QoS and fault-tolerance in distributed real-time systems

Participants : Florian Greff, Laurent Ciarletta, Arnauld Samama [Thales TRT] , Eric Dujardin [Thales TRT] , Ye-Qiong Song [contact] .

The QoS must be guaranteed when dealing with real-time distributed systems interconnected by a network. Not only task schedulability in processors, but also message schedulability in networks should be analysed for validating the system design. Fault-tolerance is another critical issue that one must take into account. In collaboration with Thales TRT industrial partner as part of a CIFRE PhD work, we started a study on the real-time dependability of distributed multi-criticity systems interconnected by an embedded mesh network (RapidIO). For easing the QoS specification at the higher level, DDS middleware is used. We postulate that enhancing QoS for real-time applications entails the development of a cross-layer support of high-level requirements, thus requiring a deep knowledge of the underlying networks. This year, we proposed and implemented a new simulation/emulation/experimentation framework called ERICA, for designing such a feature. ERICA integrates both a network simulator (Ptolemy) and an actual hardware network to allow implementation and evaluation of different QoS-guaranteeing mechanisms. It also supports real-software-in-the-loop, i.e. running of real applications and middleware over these networks. Each component can evolve separately or together in a symbiotic manner, also making teamwork more flexible [68] , [33] .

Wireless sensor and actuator networks

Participants : Lei Mo, Xiufang Shi [Zhejiang University] , Jiming Chen [Zhejiang University] , Ye-Qiong Song [contact] .

Wireless sensor and actuator networks provide a key technology for fully interacting within a CPS (Cyber-Physical System). However, the introduction of the mobile actuator nodes in a network rises some new challenging issues. In this context, we addressed two important issues: the multiple target tracking using both fixed and mobile sensors and the optimal scheduling of mobile wireless energy chargers (actuators) for fixed sensor nodes.

In our work, the data association problem in multiple target tracking is investigated. To reduce the computational complexity of traditional Joint Probabilistic Data Association (JPDA) algorithm, a modified JPDA algorithm is proposed to execute data association in multiple target tracking by utilizing the information of occlusion conditions, which is identified by a three-step algorithm. Simulation results show that the proposed algorithm has good tracking performance but low computational complexity [45] .

We also investigated the multiple mobile chargers coordination problem that is minimizing the energy expenditure of the mobile chargers while guaranteeing the perpetual operation of the wireless sensor network. We formulated this problem as a mixed-integer linear program (MILP). To solve this problem efficiently, we proposed a novel decentralized method which is based on Benders decomposition. The multiple mobile chargers coordination problem is then decomposed into a master problem (MP) and a slave problem (SP), with the MP for mobile chargers scheduling and the SP for mobile chargers moving and charging time allocation. The convergence of proposed method is analyzed theoretically. Simulation results demonstrated the effectiveness and scalability of the proposed method [38] .

Big Data-oriented networking

Participants : Jérôme François [contact] , Lautaro Dolberg [University of Luxembourg] , Thomas Engel [University of Luxembourg] , Raouf Boutaba [University of Waterloo] , Reaz Ahmed [University of Waterloo] , Shihabur Rahman Chowdhury [University of Waterloo] .

Performances of Big Data applications are tightly coupled with the performance of the network in supporting large data transfers. Deploying high-performance networks in data centers is thus vital but configuration and performance management as well as the usage of the network are of paramount importance. We thus surveyed helpful approaches in a book chapter [55] . This chapter starts by discussing the problem of virtual machine placement and its solutions considering the underlying network topology. It then provides an analysis of alternative topologies highlighting their advantages from the perspective of Big Data applications needs. In this context, different routing and flow scheduling algorithms are discussed in terms of their potential for using the network most efficiently. In particular, Software-Defined Networking relying on centralized control and the ability to leverage global knowledge about the network state is propounded as a promising approach for efficient support of Big Data applications.