Retour au site du LAAS-CNRS

Abstract

Monitoring and analysis of QoS are crucial steps for the provisioning of self-healing web services and for managing web service-based distributed interactive applications. Dealing with these issues becomes even more challenging when applications are dynamically built by composition of distributed services involving different service providers. In this case, assuming access to the internal logic and its implementation within the composed web services is not realistic. In this paper, we propose an architectural framework for monitoring and analysis of QoS driven by models for QoS analysis. This framework has been implemented and experimented for the web service technology within the European WS-DIAMOND1 project. We consider the general context where only SOAP messages between web services are monitored. The main novelty of our approach is, on the one hand, to provide a generic application-independent framework. On the other hand, we provide models allowing QoS deficiencies to be detected and considered as an indicator of the health degradation of the monitored web services.

Mots-Clés / Keywords

Abstract

Monitoring and analysis of QoS are crucial steps for the provisioning of self-healing web services and for managing web service-based distributed interactive applications. Dealing with these issues becomes even more challenging when applications are dynamically built by composition of distributed services involving different service providers. In this case, assuming access to the internal logic and its implementation within the composed web services is not realistic. In this paper, we propose an architectural framework for monitoring and analysis of QoS driven by models for QoS analysis. This framework has been implemented and experimented for the web service technology within the European WS-DIAMOND 1 project. We consider the general context where only SOAP messages between web services are monitored. The main novelty of our approach is, on the one hand, to provide a generic application-independent framework. On the other hand, we provide models allowing QoS deficiencies to be detected and considered as an indicator of the health degradation of the monitored web services.

Abstract

This paper presents a framework of architecture-centric models to support the automated and adaptive deployment of communication services for QoS-enabled end-to-end group communication systems. The Transport level (TCP, UDP level) and the above messaging Middleware level are considered as the two communication levels targeted by the QoS-driven adaptation process. Application to crisis management systems (CMS) is considered as a case study from the more general domain to which our results apply: cooperative activity support systems. The adaptation rules rely on graph matching and graph rewriting. The adaptation enactment is based on the dynamic composition of micro-protocols at the Transport level and on the dynamic binding of software components and services at the Middleware level. The deployment model is used as a central feature of service provisioning. The influence of the cooperation and the communication contexts is expressed and maintained consistent by automated graph-based model refinement and transformation.

Mots-Clés / Keywords

Abstract

Future network environments are likely to be used by cooperative applications. Indeed, the recent advent of peer to peer systems motivates this assumption. In this paper, we present a method that relies on graphs and graph grammar productions to automatically refine a high level service interactions representation of an activity into a deployment topology at the Middleware and the Transport level. At both levels, algorithms for service optimization are presented. The different models and algorithms are implemented in a case study of CMS-like operations for crisis management.

Mots-Clés / Keywords

Abstract

Future network environments are likely to be used by cooperative applications. Indeed, the recent advent of peer to peer systems motivates this assumption. In this paper, we present a method that relies on graphs and graph grammar productions to automatically refine a high level service interactions representation of an activity into a deployment topology at the Middleware and the Transport level. At both levels, algorithms for service optimization are presented. The different models and algorithms are implemented in a case study of CMS-like operations for crisis management.

Mots-Clés / Keywords

Résumé

Le monitoring et la gestion de la QdS sont des étapes cruciales pour l'approvisionnement des applications interactives distribuées. Manier ces aspects, pour des applications construites dynamiquement à base de Service Web provenant de sources différentes, devient un challenge à relever. Dans ce cas, l'accès à la logique interne et à l'implémentation des Services Web composés, au cours de l'exécution, n'est pas réaliste. Dans ce papier, nous présentons une approche d'autoadaptabilité capable de gérer les applications distribuées à base de Service Web. Ce travail a été réalisé dans le cadre du projet WS-DIAMOND1. Il couvre le cycle d'adaptation en cours de l'exécution, comportant le monitoring, l'analyse des QdS, et la reconfiguration des services à QdS dégradée.

Abstract

Handling context-aware dynamically adaptable architectures contributes to the design of self-configuring software systems. Dealing with such a problem for communicating systems is even more challenging since adaptation should address simultaneously the different communication levels. This is necessary for handling both unpredictable changes in the low level constraints and predictable evolutions in the high level requirements. In this paper, we address this problem by providing a model-based, rule-oriented approach that supports the adaptation process based on a run-time transformation of the system architecture. Such architecture may represent the different possible service compositions and the associated architectural configurations. We consider the multi-level models of the communicating system architecture. We consider the intra-level architecture transformations as the elementary adaptation actions. We handle consistently the related inter-level adaptation actions by considering additional architectural relationships viewing the lower level architecture as a refinement or a mapping of the upper level.We provide the algorithms characterizing the multi-level architecture-based adaptation process. We then develop a rule-oriented implementation using graph grammar and handling architectural transformations as graph rewriting rules. We consider Crisis Management Systems (CMS) as a case study from the more general group communication systems to which our results apply.

Mots-Clés / Keywords