Johann Bourcier

Dissemination of reconfiguration policies on mesh networks

Component-based platforms are widely used to develop and deploy distributed pervasive system that exhibit a high degree of dynamicity, concurrency, distribution, heterogeneity, and volatility. This paper deals with the problem of ensuring safe yet efficient dynamic adaptation in a distributed and volatile environment. Most current platforms provide capabilities for dynamic local adaptation to adapt these systems to their evolving execution context, but are still limited in their ability to handle distributed adaptations. Thus, a remaining challenge is to safely propagate reconfiguration policies of component-based systems to ensure consistency of the architecture configuration models over a dynamic and distributed system. In this paper we implement a specific algorithm relying on the models at runtime paradigm to manage platform independent models of the current system architecture and its deployed configuration, and to propagate reconfiguration policies. We evaluate a combination of gossip-based algorithms and vector clock techniques that are able to propagate these policies safely in order to preserve consistency of architecture configuration models among all computation nodes of the system. This evaluation is done with a test-bed system running on a large size grid network.

A prediction-driven adaptation approach for self-adaptive sensor networks

Engineering self-adaptive software in unpredictable environments such as pervasive systems, where networks ability, remaining battery power and environmental conditions may vary over the lifetime of the system is a very challenging task. Many current software engineering approaches leverage run-time architectural models to ease the design of the autonomic control loop of these self-adaptive systems. While these approaches perform well in reacting to various evolutions of the runtime environment, implementations based on reactive paradigms have a limited ability to anticipate problems, leading to transient unavailability of the system, useless costly adaptations, or resources waste. In this paper, we follow a proactive self-adaptation approach that aims at overcoming the limitation of reactive approaches. Based on predictive analysis of internal and external context information, our approach regulates new architecture reconfigurations and deploys them using models at runtime. We have evaluated our approach on a case study where we combined hourly temperature readings provided by National Climatic Data Center (NCDC) with fire reports from Moderate Resolution Imaging Spectroradiometer (MODIS) and simulated the behavior of multiple systems. The results confirm that our proactive approach outperforms a typical reactive system in scenarios with seasonal behavior.

Autonomic iPOJO: Towards Self-Managing Middleware for Ubiquitous Systems

The recent proliferation of ever smaller and smarter electronic devices, combined with the introduction of wireless communication and mobile software technologies enables the construction of a large variety of pervasive applications, such as home supervision and alarm systems. The inherent complexity of such applications along with their nonexpert clientele raises the necessity for autonomic management solutions. Nonetheless, such solutions remain difficult to conceive, as they must deal with the increased volatility, heterogeneity and distribution of the pervasive domain, while ensuring stringent performance and dependability requirements. This paper proposes that reusable support for autonomic management solutions be provided by middleware platforms, along with already existing middleware services, such as security and transactions. Following this approach, a service oriented component platform, iPOJO, was extended with elementary autonomic management capabilities. These include monitoring and effector touch points, as well as embedded autonomic management functions, such as service dependency management. IPOJO is an open source Apache project and has been successfully employed to implement several research projects in the pervasive domain. This paper presents iPOJOpsilas extension with reusable autonomic management middleware services.

Multitier Diversification in Web-Based Software Applications

Web application development benefits massively from modular architectures and reuse. This excellent software engineering practice is also the source of a new form of monoculture in application-level co de, which creates a potential risk for dependability. Researchers propose using software diversification in multiple components of Web applications to reconcile the tension between reuse and dependability. This article identifies key enablers for the effective diversification of software, especially at the application-code level. Its possible to combine different software diversification strategies, from deploying different vendor solutions to fine-grained code transformations, to provide different forms of protection.

Dependability-driven runtime management of service oriented architectures

Software systems are becoming more and more complex due to the integration of large scale distributed entities and the continuous evolution of these new infrastructures. All these systems are progressively integrated in our daily environment and their increasing importance have raised a dependability issue. While Service oriented architecture is providing a good level of abstraction to deal with the complexity and heterogeneity of these new infrastructures, current approaches are limited in their ability to monitor and ensure the system dependability. In this paper, we propose a framework for the autonomic management of service oriented application based on a dependability objective. Our framework proposes a novel approach which leverages peer to peer evaluation of service providers to assess the system dependability. Based on this evaluation, we propose various strategies to dynamically adapt the system to maintain the dependability level of the system to the desired objective.

Scapegoat: An Adaptive Monitoring Framework for Component-Based Systems

Modern component frameworks support continuous deployment and simultaneous execution of multiple software components on top of the same virtual machine. However, isolation between the various components is limited. A faulty version of any one of the software components can compromise the whole system by consuming all available resources. In this paper, we address the problem of efficiently identifying faulty software components running simultaneously in a single virtual machine. Current solutions that perform permanent and extensive monitoring to detect anomalies induce high overhead on the system, and can, by themselves, make the system unstable. In this paper we present an optimistic adaptive monitoring system to determine the faulty components of an application. Suspected components are finely instrumented for deeper analysis by the monitoring system, but only when required. Unsuspected components are left untouched and execute normally. Thus, we perform localized just-in-time monitoring that decreases the accumulated overhead of the monitoring system. We evaluate our approach against a state-of-the-art monitoring system and show that our technique correctly detects faulty components, while reducing overhead by an average of 80%.

Polymer: A model-driven approach for simpler, safer, and evolutive multi-objective optimization development

Multi-Objective Evolutionary Algorithms (MOEAs) have been successfully used to optimize various domains such as finance, science, engineering, logistics and software engineering. Nevertheless, MOEAs are still very complex to apply and require detailed knowledge about problem encoding and mutation operators to obtain an effective implementation. Software engineering paradigms such as domain-driven design aim to tackle this complexity by allowing domain experts to focus on domain logic over technical details. Similarly, in order to handle MOEA complexity, we propose an approach, using model-driven software engineering (MDE) techniques, to define fitness functions and mutation operators without MOEA encoding knowledge. Integrated into an open source modelling framework, our approach can significantly simplify development and maintenance of multi-objective optimizations. By leveraging modeling methods, our approach allows reusable optimizations and seamlessly connects MOEA and MDE paradigms. We evaluate our approach on a cloud case study and show its suitability in terms of i) complexity to implement an MOO problem, ii) complexity to adapt (maintain) this implementation caused by changes in the domain model and/or optimization goals, and iii) show that the efficiency and effectiveness of our approach remains comparable to ad-hoc implementations.

Surrogate-assisted optimisation of composite applications in mobile ad hoc networks

Infrastructure-less mobile ad-hoc networks enable the development of collaborative pervasive applications. Within such dynamic networks, collaboration between devices can be realised through service-orientation by abstracting device resources as services. Recently, a framework for QoS-aware service composition has been introduced which takes into account a spectrum of orchestration patterns, and enables compositions of a better QoS than traditional centralised orchestration approaches. In this paper, we focus on the automated exploration of trade-off compositions within the search space defined by this flexible composition model. For the studied problem, the evaluation of the fitness functions guiding the search process is computationally expensive because it either involves a high-fidelity simulation or actually requires calling the composite service. To overcome this limitation, we have developed efficient surrogate models for estimating the QoS metrics of a candidate solution during the search. Our experimental results show that the use of surrogates can produce solutions with good convergence and diversity properties at a much lower computational effort.

Optimizing Multi-objective Evolutionary Algorithms to Enable Quality-Aware Software Provisioning

Elasticity is a key feature for cloud infrastructures to continuously align allocated computational resources to evolving hosted software needs. This is often achieved by relaxing quality criteria, for instance security or privacy because quality criteria are often conflicting with performance. As an example, software replication could improve scalability and uptime while decreasing privacy by creating more potential leakage points. The conciliation of these conflicting objectives has to be achieved by exhibiting trade-offs. Multi-Objective Evolutionary Algorithms (MOEAs) have shown to be suitable candidates to find these trade-offs and have been even applied for cloud architecture optimizations. Still though, their runtime efficiency limits the widespread adoption of such algorithms in cloud engines, and thus the consideration of quality criteria in clouds. Indeed MOEAs produce many dead-born solutions because of the Darwinian inspired natural selection, which results in a resources wastage. To tackle MOEAs efficiency issues, we apply a process similar to modern biology. We choose specific artificial mutations by anticipating the optimization effect on the solutions instead of relying on the randomness of natural selection. This paper introduces the Sputnik algorithm, which leverages the past history of actions to enhance optimization processes such as cloud elasticity engines. We integrate Sputnik in a cloud elasticity engine, dealing with performance and quality criteria, and demonstrate significant performance improvement, meeting the runtime requirements of cloud optimization.

Exploring Optimal Service Compositions in Highly Heterogeneous and Dynamic Service-Based Systems

Dynamic and heterogeneous service-oriented systems present challenges when developing composite applications that exhibit specified quality properties. Resource heterogeneity, mobility, and a large number of spatially distributed service providers complicate the process of composing complex applications with specified QoS requirements. This PhD project aims at enabling the efficient run-time generation of service compositions that share functionality, but differ in their trade-offs between multiple competing and conflicting quality objectives such as application response time, availability and consumption of resources. In this paper we present a research roadmap towards an approach for flexible service composition in dynamic and heterogeneous environments.