Apto: A MDD-Based Generic Framework for Context-Aware Deeply Adaptive Service-Based Processes

Abstract

Abstract-Context-awareness and adaptability are important and desirable properties of service-based processes designed to provide personalized services. Most of the existing approaches focus on the adaptation at the process instance level [1] which involves extending the standard Business Process Execution Language (BPEL) and its engine or creating their own process languages (e.g. [2]). However, the approach proposed here aims to apply an adaptation to processes modeled or developed without any adaptation possibility in mind and independently of specific usage contexts. In addition, most of the existing approaches tackle the adaptation on the process instance or definition levels by explicitly specifying some form of variation points. This, however, leads to a contradiction between how the architect logically views and interprets differences in the process family and the actual modeling constructs through which the logical differences must be expressed. We introduce the notion of an evolution fragment and evolution primitive to capture the variability in a more logical and independent way. Finally, the proposed approach intends to support the viewpoint of context-aware adaptation as a crosscutting concern with respect to the core "business logic" of the process. In this way, the design of the process core can be decoupled from the design of the adaptation logic. To this end, we leverage ideas from the domain of model-driven development (MDD) and generative programming.

Comprehensive Variability Modeling and Management for Customizable Process-Based Service Compositions

“…From the variability descriptor, a separate BPEL process is generated to facilitate consumers in deciding variants to be bound. In addition, the approach in [12] defines notations based on Eclipse Modeling Framework (EMF) [7] for representing BPEL process models with modifiable elements. Separate context model and linkage model help to decide what, when, and how to modify the process (e.g.…”

Enabling Runtime Evolution of Context-Aware Adaptive Services

“…Part of the evolution script to evolve the travel guide service //Adding new roles organizer.addNewRole("FR4", "AttractionsLocator"); //Adding new contracts organizer.addNewContract("FC3", "User-AttractionsLocator ", "User and attractions finder contract " , "FR1", "FR4"); organizer.addNewTerm("I1", "FindAttractions", "void", "String#Location", "AtoB", "FC3"); //Changing the already defined contracts organizer.addNewTerm("I3", "NotifySpeed","void", "int#VehilceSPeed*int#SpeedLimit", "AtoB", "FC2"); //Adding new facts organizer.addNewFact("TrafficInfo", "SpeedLimit"); //Changing roles players binding organizer.changePlayerBinding("FR4","http://localhost:8080/axis2 /services/AttractionsFinder "); //Update defined processes organizer.addProcess("P2","FindAttractions","AttractionsPrefRequest", "AttractionsViewd | AttractionsSelected"); organizer.addBehaviorTerm("P2_BI1_getAttractionsPref"); organizer.addTaskToBehavior("P2_BI1_getAttractionsPref","CR2. getAttractionsPrefPref_Task", "AttractionsPrefRequested"," AttractionsPrefPrefAvailable"); organizer.addBehaviorToProces("P2","P2_BI1_ getAttractionsPref "); Second, to execute the generated evolution script, we have used Javaassist 6 to create a Java method on-the-fly. This method has a reference to the running service, and the Java code that corresponds to the differences between the service's new and old models (i.e.…”

“…In the past few years, a number of approaches have been introduced to enable a service's runtime adaptation to cope with context changes (e.g. [4][5][6]). However, these approaches only support the service adaptation in response to changes that have been anticipated during its development time.…”

Dynamic Evolution of Service Architecture in Mobile Cloud Applications

“…As a proof-of-concept we implemented a prototype tool (Apto) for the service evolution with process variant generation technology (Jaroucheh et al, 2010). We present here a case study for a tourism service application running in a multi-touch multiuser table that provides "intelligent" offering of tourism information.…”