Recent Results

Enhancing Planning-Based Adaptation Middleware with Support for Dependability: a Case Study — by Romain Rouvoy — last modified 2008-07-21 14:01

Recent evolutions of mobile devices have opened up for new opportunities for building advanced mobile applications. In particular, these applications are capable of discovering and exploiting software and hardware resources that are made available in their environment. A possible approach for supporting these ubiquitous interactions consists in adapting the mobile application to reflect the functionalities that are provided by the environment. However, these approaches often fail in offering a sufficient degree of resilience to potential device, network, and software failures, which are particularly frequent in ubiquitous environments. Therefore, the contribution of this paper is to integrate the dependability concern in the process of mobile applications adaptation. In particular, we propose to reflect dependability mechanisms as alternative configurations for a given application. This reflection allows the planning-based adaptation middleware to automatically decide, based on contextual information, to enable the support for dependability or not.

Developing self-adaptive mobile applications and services with Separation-of-Concerns — by Nearchos Paspallis — last modified 2008-12-03 12:04

Modern trends in mobile computing have inspired a new paradigm for the development of adaptive applications and services. Because of the variability which characterizes the context of such environments, it is important that the used software is developed so that its extra-functional behavior is adapted at runtime with the aim of dynamically optimizing the end-user experience. This chapter proposes a novel approach for the development of adaptive, mobile applications and services using the Separation-of-Concerns paradigm, as it was studied and designed by the Mobility and Adaptation Enabling Middleware (MADAM) project. The proposed approach specifies a set of steps for developing adaptive applications which can be automatically managed and dynamically adapted by an underlying middleware layer. Most notably, this approach treats the adaptive behavior of the corresponding application or service as a crosscutting concern of which the specification can be separated from the implementation of its functional logic.

Middleware for Dependable Adaptation — by Romain Rouvoy — last modified 2008-07-21 14:00

Ubiquitous computing systems are characterised by high variability of their execution conditions. In particular, the user devices that support these applications require continuous software adaptations to face the changes in the user environment (e.g., physical location, network connectivity, energy consumption), while maximising the application usability with regards to the user needs. Adaptation systems, such as MADAM or QuA, are extensively employing user-centered and planning-based adaptation mechanisms that provide a middleware support to component-based applications running on a mobile device. These adaptation systems collect contextual information related to user environment and compute seamlessly, when needed, the best con¯guration for the distributed application. However, the state-of-the-art adaptation methods and technologies do not consider possible device, network, and software failures; they may cease being functional in face of such failures. In fact, most approaches to adaptation are centralized: they assume an existence of a single entity that is highly available and that possesses complete knowledge of all distributed components in the entire system. This work, in the frame of the MUSIC European project, aims at extending the existing techniques for distributed adaptation with ability to handle dynamic changes in the composition of network nodes, system components and their characteristics as well as network topology. In order to provide this churn resilience, we set two objectives: firstly, our planning algorithm provides a degree of redundancy in the output composition of components. This enhances the ability of the system itself to cope with failure of non-dependable components. Our second objective is to make the planning mechanism itself fault-tolerant. The first and foremost element of our solution is de¯ning the precise level of guarantees (also called reliability QoS) in presence of failures and proving feasibility of these guarantees in a variety of practical failure models. We pursue the goal of graceful degradation of guarantees in accordance to the type of allowed failures: while it may be plausible to target perfect planning for small-scale systems with benign failures, the mechanisms required to cope with byzantine failures are generally costly, and they only allow for approximate and partial planning. In order to provide multiple levels of guarantees, we need to deploy a variety of mechanisms that achieve adaptation planning in a coordinated fashion. The mechanisms belong to two categories: i) consistent replication techniques based on reliable group communication, and ii) strong agreement protocols, such as Paxos, prove invaluable for providing strong guarantees in small- to medium-scale systems. The exact replication mechanism matches both the underlying topology and reliability QoS required by the application. In order to achieve weaker guarantees in large-scale heterogeneous systems with inherent high churn, a different type of mechanism is required, namely, fault-tolerant application-level overlays combined with gossip-based routing.

Brokering planning metadata in a P2P environment — by Romain Rouvoy — last modified 2008-07-21 14:04

In self-adaptive systems, metadata about resources in the system (e.g., services, nodes) must be dynamically published, updated, and discarded. Current adaptive middleware approaches use statically configured, centralised repositories for storing and retrieving of such metadata. In peer-to-peer (P2P) environments, we can not assume the existence of server nodes that are always available for hosting such centralised services. However, the metadata repository is the keystone of the adaptation middleware and the consistency of adaptations relies on its reliability. To address this limitation in our QuA planning-based adaptation middleware, we introduce a P2P broker, which is a metadata advertisement service based on P2P technology. This P2P broker can be plugged into the QuA middleware to support the construction of self-adaptive applications in a P2P environment. We use a structured P2P protocol that distributes the service metadata over a set of nodes based on service type and property information. The P2P broker is therefore capable of handling node failures by providing replication of the metadata. We present a working prototype of the P2P broker as well as results from initial experiments. These results show that the metadata distributes well over the nodes in the network, thus enabling scalability and robustness to node failures.

A Context Query Language for Pervasive Computing Environments — by Nearchos Paspallis — last modified 2008-03-08 12:36

This paper identifies requirements for querying and accessing context information in mobile and pervasive computing environments. Furthermore it studies existing query languages and it is shows that they satisfy only a subset of these requirements or cover some of them only to a limited extend. Therefore, we present a new context query language to overcome these shortcomings. It improves the state of the art in several respects as it explicitly addresses heterogeneous representations of context information, definition of complex filtering mechanisms, elaborate aggregation functions and ontology integration, all in one language.

A Comprehensive Context Modeling Framework for Pervasive Computing Systems — by Nearchos Paspallis — last modified 2008-07-18 20:12

Context management in pervasive computing environments must reflect the specific characteristics of these environments, e.g. distribution, mobility, resource-constrained devices, heterogeneity of context sources, context reasoning, etc. Although a number of context models have been presented in the literature, none of them supports all of these requirements to a sufficient extent at the same time. In this paper, we present a comprehensive and integrated approach for context modeling in pervasive computing environments. It combines the advantages of existing approaches and addresses the need for supporting effective software development. The proposed context model follows an ontology-based approach and has three layers of abstraction, i.e. conceptual layer, exchange layer, and functional layer. This layered approach facilitates a model-driven development of context-aware applications. Throughout the paper we compare our solution with the related work in order to clearly demonstrate why we needed to develop a new context management framework and where we have adopted existing ideas.

An Optimization of Context Sharing for Self-adaptive Mobile Applications — by Nearchos Paspallis — last modified 2008-03-08 12:26

Because of the high potential of mobile and pervasive computing systems, there is an ongoing trend in developing applications exhibiting context awareness and adaptive behavior. While context awareness guarantees that the applications are aware of both their context and their own state, dynamic adaptivity enables them to react on their knowledge about it and presumably optimize their offered services. However, because in pervasive computing environments there is also a need for enabling arbitrary synergies, such a behavior also requires appropriate adaptation logic algorithms which can enable distributed applications to reason on the sensed context and dynamically decide on the most appropriate adaptations. This paper discusses how utility function-based approaches can use context-awareness for that and, additionally, it shows how the decision-making process is improved with respect to both performance and resource consumption, by using a more intelligent approach.

D4.2 System Design of the MUSIC Architecture — by Alessandro Mamelli — last modified 2009-07-07 14:16

MUSIC is a focused initiative with the goal of designing, developing and distributing an open-source software framework, which enables developers to create context-aware mobile applications that dynamically and seamlessly adapt to changes in the user and execution context. MUSIC aims at defining a methodology to support the design and modelling of self-adaptive applications, and at providing a distributed architecture and middleware and tools to support the efficient execution of these applications in ubiquitous environments. The goal of Work Package 4 “Architecture” is designing the MUSIC architecture. It targets the above mentioned mobile and ubiquitous environments which are highly distributed, dynamic and heterogeneous. The purpose of this document is to define the initial design of the MUSIC architecture. It specifies the functionalities of major architectural components and their relationship, and takes into account initial results from Work Packages 1 (“Adaptation”) and 2 (“Context”) and the requirements that have been identified in the Deliverable D4.1. The initial design focuses on the notion of services, distributed architecture and service discovery. It lays the foundation to allow the architecture design to evolve towards service-oriented self-adaptations and selfadaptability of the MUSIC architecture. The initial design of the MUSIC architecture will be used as input for the Work Packages 5 (“Middleware”) and 7 (“Tools”) to develop the initial version of the MUSIC middleware and tools.

D8.1 - D8.5 Field Trial Scenario Framework and Scenarios — by Nicoletta Salis — last modified 2008-02-25 15:10

One of the main objectives of MUSIC is to develop, using the Project technical platform, trial services based on a set of challenging application scenarios with real market potential, interesting promotion and exploitation and as sources of technical requirements. This document is the result of the work performed in Work Package 8 “Scenarios”, and aims at providing the guidelines for the design of the MUSIC innovative applications and the description of the related field trials, defining the involved entities (persons or machines) and the links among them. Starting from a general definition of context and self-adaptation leads to a selection of some high-level parameters necessary to build the view of the real-time user situation and of the needed adaptation level; these parameters are the basis to build: an overview of the service scenarios which can be the most important ones in our perspective a general service architecture showing the relationships among the main actors, as a kind of framework for the different specific architectures of the field trials. For each field trial, inserted in the appropriate service scenario, a good description is given, focusing on: giving a general description of the objectives of the trial and of its characteristics (where, user target, which physical environment, which constraints …) highlighting the main technical features of the service describing the current implementation of this use case (if it already exists) or the examples of the state of the art pointing out the differences and the advantages of the introduction of MUSIC functionalities specifying the technical requirements in order to give feedback to the Activity Groups A and B of the Project Keywords: Scenarios, Services, Field Trials, Context-Awareness, Self-Adaptation URL: http://www.ist-music.eu/docs/D8.1-D8.5.pdf

D1.2 Initial research results on mechanisms and planning algorithms for self-adaptation — by Romain Rouvoy — last modified 2009-01-24 11:00

This deliverable is the second published result of the 1st Work Package of the MUSIC project. The main objective of this deliverable is to document the initial results on mechanisms and planning algorithms for self-adaptation developed by the MUSIC partners. Additionally, this deliverable aims to document the main requirements that emerge during the development and deployment of commercial, adaptive, mobile applications, targeting ubiquitous computing environments. In the MUSIC project, we are primarily concerned with developing methods and techniques for achieving adaptable software. The main target of the desired adaptation concepts, management system and mechanisms is to support both the development and run-time management of software systems that are capable of being adapted to varying contextual situations in ubiquitous computing environments. We are aiming for an adaptation system that clearly separates and externalizes the logic of adaptation mechanisms and policies from the code defining the application logic. This will enable the provision of adaptive behaviour as generic and reusable services relieving the application programmer from much of the complex tasks of encoding adaptation management and mechanisms. In order to facilitate adaptation services as indicated above, they must be based upon appropriate concepts modelling adaptability of component-based distributed applications. Furthermore, corresponding dependable adaptation mechanisms that can be applied at run-time must support design time adaptation specification elements. The adaptation services must also include policies for deciding when and how to adapt, as well as policies for undertaking the corresponding reconfiguration in a safe and timely manner. In this respect, this document also studies the state of the art in these areas. Additionally, as the MUSIC project aims at providing an implementation of a middleware of which the adaptation management will be a crucial part, this document tries to collect a set of requirements that will be needed by the applications using the middleware as well as the application developers who use the provided models and tools to create them. This document is composed of two parts, one for each of the two main activities we have focused on. The first part, from Chapters 2 to 5, refers to the research activities in the domain of service-oriented architecture. The second part of the document, from Chapters 6 to 9, focuses on the issues related to the scalability and the performances of the planning middleware.

D7.1 Modelling Tools — by Jorge Lorenzo — last modified 2008-02-22 09:24

MUSIC is a focused initiative that develops a comprehensive open-source software development framework that facilitates the development of self-adapting, reconfigurable software that seamlessly adapts to the highly dynamic user and execution context, and maintains a high level usefulness across context changes. Context-aware applications are capable of exploiting knowledge of external operating conditions, and they are self-adaptive by adapting at runtime to varying contexts, like changing user needs and operating environments. MUSIC provides a design methodology and distributed system architecture for the design and implementation of self-adapting applications in ubiquitous computing environments. This is complemented with enhanced modelling languages for the specification of context dependencies and adaptation capabilities, supported by model specification, validation and simulation tools. This platform is also used to develop trial services, based on a set of challenging application scenarios with real market potential, having a central role: as sources of requirements, to assess technical adequacy of the results, and to promote the results. This deliverable is a result of the work performed in the Work Package 7 of MUSIC, and aims at providing the consortium with recommendation of a tool for UML modelling and also another one for ontology modelling, after having studied most of the available possibilities in both the market and especially in the open source community. Comparative criteria are established in this document and the results obtained by each of the tools are presented. Keywords: Modelling tool, UML, Ontology, criteria for tools evaluation

D5.1 Design of the adaptation middleware — by Jorge Lorenzo — last modified 2008-02-22 09:23

MUSIC is a focused initiative that will develop a comprehensive open-source software development framework that facilitates the development of self-adapting, reconfigurable software that seamlessly adapts to the highly dynamic user and execution context, and maintains a high level usefulness across context changes. Context-aware applications are capable of exploiting knowledge of external operating conditions, and they are self-adaptive if they adapt at runtime to varying contexts, like changing user needs and operating environments. Work Package 5 is focused on the detailed design and implementation of the MUSIC middleware which in turn is based on the MADAM middleware. Due to the incremental project model we are following in MUSIC, most deliverables will be snapshots describing results achieved so far, which will often be preliminary results to be superseded by more complete and mature results described in later deliverables. The MUSIC architecture was not mature enough during the first stage of the project in order to start the detailed design of the MUSIC middleware. As a result, deliverable 5.1 has addressed two different purposes during the first 12 months:         • The state of the art about different runtime platforms (software and hardware) as well as development environments to build the MUSIC middleware. This research aims to get the most suitable tools to develop this middleware throughout the rest of the project.     • The enhancement of the existing MADAM middleware with improvements of different types: to correct bugs, to get better performance, to extend it to a Service-Oriented approach and so on. All necessary information about MADAM adaptation improvements is provided in this document while context-aware improvements are documented in D5.3. Keywords: Middleware, development environment, adaptation, design, requirements, MADAM Improvements, software platform, hardware platform.

D5.3 Design of the context middleware — by Massimo Valla — last modified 2008-02-22 11:47

MUSIC is a focused initiative that will develop a comprehensive open-source software development framework that facilitates the development of self-adapting, reconfigurable software that seamlessly adapts to the highly dynamic user and execution context, and maintains a high level usefulness across context changes. Context-aware applications are capable of exploiting knowledge of external operating conditions, and they are self-adaptive if they adapt at runtime to varying contexts, like changing user needs and operating environments. Within MUSIC, the Context Middleware is the software component responsible for the acquisition of context information from sensors, allowing access to existing context and enabling efficient context distribution among nodes. MUSIC approaches software development in an initial phase and four development cycles. During the initial phase, the MUSIC middleware architecture was not mature enough in order to start the detailed design of the MUSIC Context Middleware. Therefore the scope of this WP5 deliverable focuses on an analysis of state-of-the art technologies for context-awareness, as a preliminary work to the following development cycles. Technologies are analyzed from other existing projects and products, focussing on: context sensing, context reasoning and storage, existing context-aware systems. A section is also dedicated to survey existing SOA frameworks. The deliverable also addresses some of the improvements, already implemented on the MADAM middleware, related to context management and identified by the developers of the MADAM pilots. Keywords: Middleware, Context technologies, Context-awareness, Technical survey, MADAM improvements URL http://www.ist-music.eu/MUSIC/docs/D5.3%20Design%20of%20the%20context%20middleware.pdf

D2.2 Initial research results on methods, languages, algorithms and tools to modeling and management of context — by Massimo Valla — last modified 2008-02-22 11:47

This deliverable is a result of the work done in the WP2 and aims at studying and proposing solutions to enable context awareness. To this respect, several related topics have been studied, including context sensing, monitoring, modeling, sharing, distribution and consumption. The primary aim of this work was to provide subsequent work packages (such as WP4 and WP5) with the required input for leveraging the design of the architecture and the implementation of the MUSIC middleware and development tools. The main results of the work documented in this deliverable are the design of a novel context modeling approach and context querying language, which enable advanced features for the development of context-aware software. Additional results are the design of a peer-to-peer based algorithm for distribution of context, based on the JXTA middleware, and the design and evaluation of alternative approaches for context distribution in wide-area networks. Finally, this document includes a study of user profiling issues, partly based on the work which was initialized in the previous deliverable of WP2 (i.e. D2.1) and also provides an evaluation of context architectures along with lessons for the development of the MUSIC middleware. Keywords: Context, Context-awareness, State-of-the-Art, Requirements, Adaptive systems, Middleware, Modeling, Semantics URL http://www.ist-music.eu/docs/D2.2.pdf

D6.1 Modelling Notation for Adaptive Applications in Ubiquitous Computing Environments — by Michael Wagner — last modified 2008-02-21 13:08

MUSIC is a focused initiative that will develop a comprehensive open-source software development framework that facilitates the development of self-adapting, reconfigurable software that seamlessly adapts to the highly dynamic user and execution context, typical of mobile and ubiquitous computing and maintains a high level usefulness across context changes. For this purpose, MUSIC will provide a methodology for the model-driven development of self-adapting applications integrating modelling language enhancements for the specification of context dependencies, application variability and adaptation capabilities. This Deliverable reflects the work done in WP6 for Tasks 6.1 in the first year of the project. Task 6.1 deals with the definition of a modelling notation allowing expressing the adaptation capabilities of an application. The modelling notation takes into account results and new requirements from WPs 1-5 and directly depends on the MUSIC context modelling approach and on the service-oriented discovery and binding features. UML is the first choice modelling language and the existing MADAM UML profile serves as a starting point for this notation. It is analysed and improved with regard to shortcomings revealed in the evaluation of the MADAM technology as a technology base for MUSIC and enhanced to cover the new requirements resulting from the other WPs. However, in particular with regard to context modelling and the integration in a service oriented computing environment, UML seems not to be sufficient. Therefore, based on an extensive analysis of existing modelling approaches we present the initial version of the MUSIC modelling notation incorporating ontologies and semantic web technologies in addition to UML, in order to address the new requirements. Keywords: Adaptive application, context-awareness, context modelling, model-driven development, UML Profile, service orientation, ontology, OWL, OWL-S, WSDL.

D6.2 Software Development Method for Adaptive Applications in Ubiquitous Computing Environments (initial version) — by Michael Wagner — last modified 2009-07-09 15:10

MUSIC is a focused initiative that will develop a comprehensive open-source software development framework that facilitates the development of self-adapting, reconfigurable software that seamlessly adapts to the highly dynamic user and execution context, typical of mobile and ubiquitous computing and maintains a high level usefulness across context changes. The design and implementation of correct, effective and dependable adaptive applications in the envisaged ubiquitous computing environment is certainly a great challenge. Therefore, it is an important objective of MUSIC, to complement the middleware forming the run-time infrastructure of adaptive applications with an elaborate development methodology to support developers. This Deliverable reflects the work done in WP6 for Task 6.2 in the first 18 months of the project. Task 6.2 deals with providing a methodology for the model-driven development of self-adapting applications integrating modelling language enhancements for the specification of context dependencies, application variability and adaptation capabilities. The methodology utilizes the modelling notation for adaptive applications presented in Deliverable D6.1 and shows how tools can be incorporated, in order to ease the development task. Therefore, this Deliverable has to be considered as important basis for WP7 work that provides the development tool suite (MUSIC Studio) supporting the methodology. The development methodology not only provides a step-by-step guideline to specify the application’s context dependencies, variability and domain model, but also covers the transformation of models to source code, the deployment of applications on the middleware and the testing and validation of the expected adaptation capabilities. Keywords: Adaptive application, model-driven development, methodology, modelling notation.

D3.2 Initial research results on methods for testing, tuning and validation of self-adaptative systems — by Jacqueline Floch — last modified 2008-02-21 15:32

The MUSIC project aims at developing a comprehensive open-source software development framework that facilitates the development of self-adapting, reconfigurable software that seamlessly adapts to the highly dynamic user and execution context, and maintains a high level usefulness across context changes. A leading idea in MUSIC is that context awareness and self-adaptation can be supported by generic solutions in the form of methods, languages, tools and middleware and in that way we can significantly reduce the additional complexity and effort normally associated with the implementation of such properties in software systems. The project proposes an architecture-centric approach where architecture models specified as design time are represented at runtime allowing generic middleware components to reason about and to control adaptation of applications. While the initial goal of the 3rd Work Package in the MUSIC project was to provide methods for the tuning and validation of property predictors specified in the architectural models, we have chosen to concentrate on performance properties (e.g. response times, throughput). Specifying such properties is an especially difficult task that developers are not familiar with. This deliverable investigates exploiting existing performance engineering solutions and presents a number of issues that those solutions raise in the context of MUSIC. In this deliverable, we first introduce a simple performance engineering approach that we have earlier positive experience with. We propose to further investigate that approach in MUSIC as an initial step. We identify the set of important resource parameters needed by the approach and investigate techniques for capturing such parameters. We foresee two important extensions to the method for better integration with the MUSIC framework: one is about addressing performance modelling in UML, the other about supporting component-based development. We present a survey of the research work for performance engineering with focus on those two issues. However, we do not yet discuss how we can exploitation the relevant research results in our approach. This is planned to be done in the next period of the project in accordance with the incremental project model. Keywords: Tuning, Validation, (Context) Simulation, Testing. URL D3.2 Initial research results on methods for testing, tuning and validation of self-adaptative systems

Composing Components and Services using a Planning-based Adaptation Middleware — by Romain Rouvoy — last modified 2008-07-21 14:03

Self-adaptive component-based architectures provide methods and mechanisms to support the dynamic adaptation of their structure under evolving execution context. Dynamic adaptation is particularly relevant in the domain of ubiquitous computing, which is subject to numerous unexpected changes of the execution context. In this paper, we focus on changes in the service provider landscape: business services may dynamically come and go, and their quality of service may vary. We introduce an extension of the MADAM component-based planning framework that optimizes the overall utility of applications when such changes occur. MADAM planning is based on dynamic configuration of component frameworks. The extended planning framework supports seamless configuration of component frameworks based on both local and remote components and services. In particular, components and services can be plugged in interchangeably to provide functionalities defined by the component framework. The extended planning framework is illustrated and validated on a use case scenario.

D3.1 Requirements of Methods for Testing, Tuning and Validation of Self-Adaptive Systems — by Gunnar Brataas — last modified 2008-02-21 10:15

The focus of this deliverable is tuning of the adaptations performed by the MUSIC middleware. Appropriate models must exist before tuning is possible. Some properties such as performance aspects (response times, throughput) are difficult to specify and will typically be described using analytical modelling. Simpler like properties of user interfaces are not addressed in this deliverable. To calibrate the parameters for analytical modelling we may use simulation or testing. To drive both the simulation and the testing we need a simulation of the context. This deliverable describes the state of the art in the areas of analytical modelling, simulation of the adaptation process and simulation of context, and finally also in the field of testing. Requirements in these areas are defined. Finally, this deliverable identifies the modelling of property predictors as the focus of the next work in WP3. Properties must be modelled both for individual components and for applications consisting of many components.

D4.1 Requirements on the MUSIC Architecture — by YunDing — last modified 2007-12-13 14:10

The main objective of this deliverable is to specify the requirements for the MUSIC architecture which serve as a basis for the design of the architecture. This deliverable also documents the state-of-the-art of software architectures for self-adaptive systems in mobile and ubiquitous environments.