Logo of MSCA (Marie Skłodowska-Curie Actions)

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement n° 813884.

Flag of European Union

Lowcomote logo

Scalable Management of Low-Code Artefacts

Wednesday, 14th April 2021


Times are in CEST.

9:00 - 9:05 Introduction
Davide Di Ruscio, University of L'Aquila
9:05 - 09:45 CDO - Repositories for Scalable Models
Eike Stepper
9:45 - 10:30 MDEForge: A model-as-a-service framework for multi-perspective mining of collaborative modelling repositories
Juri Di Rocco
10:30 - 10:40
10:40 - 11:20 Scalable model queries with Hawk
Antonio Garcia-Dominguez
11:20 - 12:00 Managing scalability with Modelio
Alessandra Bagnato
12:00 - 13:30
Lunch break
13:30 - 14:10 Reactive model processing using VIATRA (and the IncQuery Model Analysis suite)
Ákos Horváth
14:10 - 14:50 Managing model-to-model transformations at scale with YAMTL
Artur Boronat
14:50 - 15:05
15:05 - 15:45 Extending Declarative ATL with Incrementality and Constraints
Frédéric Jouault
15:45 - 16:25 How OpenMBEE Enables the Adoption of a Model-Based Digital Ecosystem at Boeing and the Jet Propulsion Laboratory
Myra Lattimore and Lucas Aviles
16:20 - 17:00 Discussion and wrap up

CDO - Repositories for Scalable Models

Abstract: Eclipse CDO is an open source technology that provides you with scalable model repositories. Building upon the Eclipse Modeling Framework (EMF) it transparently enables you to store, lock, and access your models at the granule of single objects. Modifications are committed from your applications to the repository server where they are properly versioned, stored in a database, and possibly propagated to other applications, which then update instantaneously. Lazy loading of single objects on demand is supported, as well as automatic memory management. In this talk I will demonstrate many of CDO’s core functions by using the CDO Explorer, a generic user interface that integrates seamlessly with Eclipse. I will also explain how EMF alone can become a barrier to working with big models and how CDO helps you to overcome this barrier. I will close with a brief listing of the typical challenges when your models grow large and what CDO offers you to help master them.


MDEForge: A model-as-a-service framework for multi-perspective mining of collaborative modelling repositories

Abstract: Model-Driven Engineering (MDE) is increasingly used across industries to abstract designs and viewpoints. Even though current modeling tools provide features that can simplify and automate many model-based development steps, obstacles to the broader adoption of MDE technologies still exist. Similar to what happens with source code development supported by version control systems, to deal with software systems’ growing complexity, it’s necessary to enforce consistent reuse and leverage the interconnection of the modeling artifacts produced and consumed during the different development phases. Besides, the availability of modeling artifacts is crucial for applying techniques like Machine Learning and Data Analytics to MDE. In this talk, first, I’ll discuss the opportunities that the availability of advanced model repositories provides. In the second part, I’ll present MDEForge: an extensible Web-based modeling platform specifically conceived to foster a community-based modeling repository, underpinning the development, analysis, and reuse of modeling artifacts.


Scalable model queries with Hawk

Abstract: In industrial modelling contexts, models may grow to have millions of interlinked elements. Using single files to store these large models is no longer practical: instead, tool developers have moved to approaches such as model fragmentation or database-backed persistency. Mature version control systems like Git can be reused to track the evolution of models that have been fragmented across many files, but it complicates querying as it may still require downloading and loading the entire memory at once. This talk will present Eclipse Hawk, a framework for providing scalable model querying. Hawk can index fragmented models into NoSQL stores, pre-computing subqueries in an incremental manner if desired, and can answer queries efficiently over its web service API. Hawk has been extended with the capability to query the history of a model, travelling back and forth in time during a single query.


Managing scalability with Modelio

Abstract: The current designs and management technologies are being stressed to their limits in terms of collaborative development, efficient management and persistence of large and complex models. Scalability has different dimensions: domains, team localizations, number of engineers, size and management of the engineering artefacts, interoperability and complexity of languages used. This presentation depicts how this challenge is tackled with the Modelio modeling tool solution.


Reactive model processing using VIATRA (and the IncQuery Model Analysis suite)

Abstract: Reactive programming provides abstractions to express event-driven applications in which data and computation dependencies are managed automatically. These applications react to events emitted by external event sources without an explicit notion of time or prior knowledge of the sequence of events.

In this talk, we present a reactive, event-driven model transformation platform based on the reactive programming paradigm that provides a unified framework based on (batch and live) queries and transformations to drive the systematic and well-founded integration of tool features in various scenarios over graph models. We will present the concept of the Event-Driven Virtual machine as realized within the open-source Viatra framework and demonstrates how reactive model transformation can easily be developed. Finally, a scaled up version of the approach based on the cloud-enabled IncQuery Model Analysis suite will also be introduced.


Managing model-to-model transformations at scale with YAMTL

Abstract: Model-to-model transformations play a useful role in low-code software development, including the specification and development of model compilers, consistency maintainers and software (co-)evolution. YAMTL is an EMF-based model-to-model transformation engine whose transformations are defined with an internal DSL of Xtend. The main strenghts of the language are its interoperability with Java, its expressiveness and its performance. In this talk, I will present the YAMTL language with use cases that showcase some of its features from a scalability perspective.


Extending Declarative ATL with Incrementality and Constraints

Abstract: Model-driven engineering techniques are now available to perform a wide variety of tasks. Moreover, when other techniques are necessary, it is often possible to translate problems to technical spaces in which they are available. Some modeling problems can thus be translated into constraint satisfaction or optimization problems in order to leverage constraint solvers. However, such translations are not always trivial, and typically require writing specific code to handle non-model artifacts. This talk presents our work on extending model transformation with support for incrementality and constraint solving. This makes it possible to declaratively constrain models, or generate constrained models. After each modification to such models, a constraint solver can fix all constraint violations if necessary. We applied this approach to ATL, starting with developing a compiler that generates scalable incremental transformation code using active operations. Then we added the possibility to specify constraints on ATL rules, which can be used to define interactively explorable model sets.

Bio: Frédéric Jouault is associate professor at ESEO, France. He received his PhD from the University of Nantes before doing a postdoc at the University of Alabama at Birmingham. His research interests involve model engineering, transformation, synchronization, and execution, as well as their application to Domain-Specific Languages (DSLs) and model-based reverse engineering. Frédéric created ATL, a DSL for model-to-model transformation.

Théo Le Calvar received his PhD in Computer Sciences from the University of Angers in France in 2019. He is currently working as a postdoc in the ERIS team at ESEO in Angers, France and in the GEODES team at Université de Montréal in Canada. His current research interests include incremental model transformation, constraint solving and modeling on the web.

How OpenMBEE Enables the Adoption of a Model-Based Digital Ecosystem at Boeing and the Jet Propulsion Laboratory

Abstract: The Open Model Based Engineering Environment (OpenMBEE) platform has become a key driver for the adoption of a model-based digital ecosystem across different organizations by connecting engineering information across various sources. OpenMBEE is an integrated set of open-source software tools which replaces traditional silos of information with consistent, traceable, and precise engineering models and documents. This talk provides an overview of the vision, motivation, and available infrastructure of OpenMBEE and its applicability to enterprise-level projects at JPL and Boeing. Lucas will share insights on how Boeing has successfully deployed OpenMBEE across its global enterprise, including a robust infrastructure as well as process and training support. Myra will discuss how OpenMBEE has been successfully integrated into JPL’s digital engineering environment which enables engineers to practice a variety of model-based techniques for systems engineering.


Other workshops: