Submit a problem

In order to provide you the best help possible to analyze and debug your problems please provide us some information about your environment.
When you submit a problem, please indicate:
- The Version, Build Id, Core Id, System and Architecture (You can get these information in the Help/About menu entry)
- The Modelio Edition (Modelio Open Source or Modelio by Modeliosoft)
- Your Operating System

× In this section, the Modelio research team publishes the results of their experiments with innovation modelling technologies in various areas including: Embedded Systems, Cyber-Physical Systems, Cloud, Big Data.

file Creator 4Clouds and MODAClouds design and runtime components

2 years 2 weeks ago - 2 years 2 weeks ago #3725 by malm
Creator 4Clouds relies on the model-driven approach developed as a Domain-Specific Language for the design and execution of applications on multiple clouds as developed during the MODAClouds project . MODAClouds implemented a set of design and runtime tools for designing and operating provider independent multi-cloud applications. Creator 4Clouds is the centre of MODAClouds design tools.



The figure above presents Creator 4Clouds and the components it integrates.
More information on the integrated tools is provided bellow.

SPACE 4Clouds
License: Apache 2.0

Documentation: here

Code repository: here

The SPACE 4Clouds tool deals with the QoS assessment of the modelled application and the optimization of the deployment configuration. It is implemented as an extension of the Palladio Bench framework and is composed by a set of Eclipse plugins. The tool takes as input a Palladio Component Model and a set of extension models with information about the usage profile, the cloud environment and the QoS constraints in XML format. In order to derive performance metrics the tool interacts with LQNS and LINE. The preferred performance engine, LINE, is implemented in MATLAB and takes as input LQNS models in the same XML format used by LQNS (For further details see MODAClouds Deliverables D5.2.2 and D5.4.2)

Venues 4Clouds
License: Apache 2.0

Documentation: here

Code repository: here

The Venues 4Clouds is a SaaS type of the application which though the set of enforced methodology, strategically defines users Business and Technical Assets, Risks and Requirement, in this case called Treatments to identify best matching sets of services for each of the architecture component. Architecture as such is read from the XML format file generated by Palladio Component Model. Once the services are selected, reviewed and re-evaluated, Venues 4Clouds generates an XML Palladio Component Model with services and providers the user has chosen.

Tower 4Clouds
License: Apache 2.0

Documentation: here

Code repository: here

The Tower 4Clouds is composed of server side components and client side components. As far as the server side is concerned, 5 components have been developed: the Monitoring Manager (MM), the Knowledge Base (KB), the Deterministic Data Analyzer (DDA) and a 2 Statistical Data Analyzer (Weka and Matlab versions). All of them are REST services, developed in Java and built with Maven, except for the Matlab SDA which is written in Matlab except for the Matlab SDA which is contained in the MATLAB Compiler Runtime. The MM is the main coordinator of the platform; the KB is an Apache Fuseki Server which contains the deployment model of the application being monitored and the data collectors configuration; the DDA is the component in charge of analysing, aggregating and evaluating constraints on monitoring data and is configured by the MM by means of CSPARQL queries; the 2 SDAs are in charge of computing statistical aggregations such as predictions or correlations.

SLA tool
License: Apache 2.0

Documentation: here

Code repository: here

The SLA tool comprises a REST server, where main features are implemented, and a set of additional helper tools. The Mediator tool is a command line application that takes as input a Palladio Component Model, the set of QoS Constraints and the set of Monitoring Rules, and generates SLA templates and SLA agreements that describe the service offering.

CloudML / Models @ runtime
License: Apache 2.0

Documentation: here

Code repository: here

CloudML is implemented with Java as programming languages and Maven as a build tool. CloudML models and metamodels are represented as plain Java objects. These models can be serialised in either JSON or XMI. The JSON and XMI codecs are based on Kotlin and the Kevoree Modeling Framework (KMF) [FouquetNainMDBPJ12]. For the IaaS level management, the provisioning and deployment engine relies on jclouds and the Flexiant Cloud Orchestrator API . For the PaaS management, the engine uses the Unified PaaS Library - Cloud4SOA - library and the Amazon Elastic Beanstalk and RDS APIs. New user interfaces can easily be added thanks to the Facade which allows manipulating deployment model and triggering a deployment. Similarly, new serialization codecs can easily be added.

Feedback loop
License: Apache 2.0

Documentation: here

Code repository: here

The Feedback loop integrates Creator 4Clouds with runtime component so that models can be updated from data obtained during the execution of actual applications.

Please Log in or Create an account to join the conversation.

Moderators: andyalebmalm
Time to create page: 0.212 seconds
^ Back to Top