Templatized Model Transformation: Enabling Reuse in Model Transformations Amogh Kavimandan and Aniruddha Gokhale Dept. of EECS, Vanderbilt University, Nashville, TN 1. Heterogeneous DRE System Development 2.1 Identifying the variabilities 3 Representative Case studies demonstrating the applicability of MTS Constraint notation blocks can be used to identify variabilities in templatized transformation e.g.: Structural variabilities, where variation emanates from dissimilarities in structural composition of application variants (e.g., model objects, their cardinalities etc.) Qualitative variabilities, where model objects may be same, but their values may be different across different application variants Specified as a special comment in transformation rules and therefore is transparent to the transformation engine Heterogeneous, consist of co-operating application subsystems, developed using several middleware technologies map requirements to configuration options Separate QoS specification mechanisms for each type of middleware – QoS configuration[2] must be performed for each of these platforms Variability is in mapping – hard to extend to other platforms! 1 2 Composed DRE system QoS configuration 2.2 Generating variability metamodel Commonalities present in these QoS mechanisms of individual platforms E.g., source/type-based filtering support, group events, push/pull-based subscription model for event-based communication Communication dialog (1) and real-time QoS configuration (2) metamodels. Each case study has certain variabilities that must be separated from the mappings MTS defines higher-order transformations that generate variability metamodel from the application variability specification in 2.1 These commonalities can be used in generalizing mapping of application QoS requirements to middleware-specific configurations Group events MTS constraint specification language Variabilities are in semantics of mechanisms, their values etc. – can be separated from the generalized QoS mappings The variability metamodel creates its model objects and attributes from those of the input and output languages of the templatized transformation In essence, moves the variability from “transformation” into a “model” Application variant-specific variabilities can be modeled as name-value pairs of object associations using the variability metamodel Commonalities in QoS configuration New techniques are needed to perform QoS configuration for composed DRE systems 2. Our Solution: Templatized Model Transformations Identification of variabilities and their specification in MTS Variability metamodel generation Table indicates what variabilities in each case study, the same variabilities are captured in MTS formal notation. Notice these variabilities are generic, i.e., no values or cardinalities are associated with them 2.3 Specializing application instances MTS higher-order transformations generate variability metamodel from the specification above It creates model objects for structural variabilities (to capture associations, cardinalities), and attributes for qualitative variabilities (to capture concrete variations in attribute values, based on individual instances) The models created form using this metamodel ultimately enable separation of variations from application transformations Both transformation and variabilities can evolve independently, new variability models (for new applications) can be added without breaking existing mappings Application developers create specialization repository in terms of models (as opposed to rules) Individual models in specialization repository denote concrete instances of application variabilities (from 2.1) MTS defines another higher-order transformation that creates specialized instances of model transformations from these repository models Templatized Model Transformations (MTS)[1]: Allow easy identification of variabilities: use a textual constraint specification language for developers to identify what changes in their application Automatically generate variability metamodels from variabilities: create a new modeling language from the system variabilities specified in 1 Synthesize a specialization repository: developers create specialization repository, i.e., instances of the variability modeling language, modeling the specific concrete variations in their applications Specialize application instances: an instance of specialization repository is used to create specialized instance of their applications MTS can be applied to any existing model transformation-based application (e.g., transformations for middleware QoS configuration mapping from the requirements) Specializing application instances using specialization repository models Variability metamodel generation for communication dialog case study [1] A. Kavimandan et. al. “Context-Sensitive Dialog Synthesis for Enterprise Workflows Using Templatized Model Transformations,” The Twelfth IEEE International Enterprise Computing Conference (EDOC 2008), München, Germany, September 15-19, 2008. [2] A. Kavimandan et. al. “Automated Middleware QoS Configuration Techniques for Distributed Real-time and Embedded Systems,” The Fourteenth IEEE Real-Time and Embedded Technology and Applications Symposium (RTAS 2008), St. Loius, MO, April 2008.