Outils pour utilisateurs

Outils du site


ubinet_internship

Différences

Ci-dessous, les différences entre deux révisions de la page.

Lien vers cette vue comparative

Les deux révisions précédentes Révision précédente
Prochaine révision
Révision précédente
ubinet_internship [2009/12/02 15:03]
tigli
ubinet_internship [2009/12/02 17:14] (Version actuelle)
tigli
Ligne 39: Ligne 39:
 intelligence. But these first ubiquitous systems were often proofs of intelligence. But these first ubiquitous systems were often proofs of
 concept with a single static configuration with a priori known devices. concept with a single static configuration with a priori known devices.
 +
 +Web service technology is a representative means of heterogeneous ​
 +system integration and communication. Process language standards, such as 
 +WS-BPEL and WS-CDL, have accelerated the usability of web services in 
 +business area. However, recently emerging web service devices in ubiquitous ​
 +environments still have a difficulty in coordinating their processes because of 
 +the limited computing power and storage. This research proposes a framework ​
 +of event-based process enactment for ubiquitous web service devices. The 
 +framework adopts P2P architecture where devices communicate with one 
 +another via web services eventing. The schema of ECA rules and messaging ​
 +protocol are presented for P2P process enactment so that service devices can 
 +interact each other and accomplish their process execution based on the ECA 
 +rules. Our proposed framework is expected to be useful in ubiquitous service ​
 +environments since it enables a scalable and light-weighted process enactment ​
 +through event-based web service technology.
 +
 +
 Today mobility of users and an increasing heterogeneity of devices Today mobility of users and an increasing heterogeneity of devices
 introduce a new significant challenge for Middleware for ubiquitous introduce a new significant challenge for Middleware for ubiquitous
Ligne 61: Ligne 78:
 (AA) to adapt reactively and independently assemblies of (AA) to adapt reactively and independently assemblies of
 components in Service Component Architecture (SCA) for example. components in Service Component Architecture (SCA) for example.
 +
  
 **Main objectives of the internship :** **Main objectives of the internship :**
 +
  
 After a short survey on reactivity in adaptive middlewares,​ the students After a short survey on reactivity in adaptive middlewares,​ the students
-of this internship will study in details the evolution of the aspects oriented approaches  +of this internship will study in details the evolution of Aspect-oriented system architecture and more particularyAspect of Assembly for Component-based architecture. 
-to Aspect-oriented system architecture and more particulary ​to Aspect of Assembly for Component-based architecture. +Then the student will highlight what features of such techniques are "​common aspect-oriented concepts" ​
-Student ​highlight what features of such techniques are "​common aspect-oriented concepts" ​+
 and what features are rather target architecture-specific (for example specific to component-based architecture for AA concept). ​ and what features are rather target architecture-specific (for example specific to component-based architecture for AA concept). ​
-Through Aspect-Oriented Modeling (AOM) he will extract such common characteristics from a perspective ​ +After this first part,he will be able to propose :  
-that is at more abstract level (i.e., target architecture-independent)+meta model of service for device and event-based process 
 +- different jointpoint models and corresponding weaving algorithms to react to appearing and disappearing services in the application. 
 +For examplejointpoint would be able to add contextual informations about the ports of the devices. In this case weaving process will  
 +be able to verify contextual constrains at runtime
  
-After this first part +The student will also consider the algorithmic costs of each weaving process, identifying a formal performance model  
 +that will allow to compare various solutions.
  
-algorithms in the different steps +To illustrate these results, some algorithms ​would be tested
-(Pointcut Matching and Weaving Mecanism) of the AA application process.  +
-They will specifically study the algorithmic costs of each of these +
-steps  and will propose a formal performance model that will allow to +
-compare different policies in the application of AA. +
-To illustrate these results, some algorithms ​will  be able to be tested+
 in real experiments on the CONTINUUM platform, in collaboration with the in real experiments on the CONTINUUM platform, in collaboration with the
 research engineer of the project. research engineer of the project.
Ligne 96: Ligne 113:
 == Duration : == == Duration : ==
 4 to 6 months with possible continuation in PhD 4 to 6 months with possible continuation in PhD
 +
 +
 +
  
  
ubinet_internship.1259762604.txt.gz · Dernière modification: 2009/12/02 15:03 par tigli