Authors:
Yiannis Verginadis, Wiktor Zolnieroiwcz, Paweł Skrzypek, Daniel Seybold, Kyriakos Kritikos, Somnath Mazumdar, Antonia Schwichtenberg, Feroz Zahid, Jörg Domaschka, Geir Horn, Ernst Gunnar Gran, Daniel Baur, Hynek Masata and Paweł Gora
Abstract
This document presents an initial system specification of the Melodic Multi-Cloud middleware platform. The document covers two important areas necessary for the subsequent development of the Melodic platform: Requirements Analysis and Technology Evaluation. The requirement analysis is essential to establish a set of initial features needed to realise an efficient middleware platform for Cross-Cloud dataintensive computing. At the same time, a complete evaluation of the available open technologies ensures that the efforts in the Melodic project are directed towards providing technology and tools that are currently beyond state-of-the-art and stateof-the-practice, and not on re-engineering solutions already available from other European and international open source projects. Based on the requirement analysis, considering both the perspective of the general data-intensive Cloud applications as well as the needs of the four Melodic use-cases, a set of functional and non-functional requirements are presented in this document. In addition, technical and business-level constraints imposed by private and public Cloud solutions are identified for the Melodic middleware platform. Moreover, a comprehensive technology evaluation comprising identification of reusable components from three identified EU projects (PaaSage, CACTOS, and PaaSword) and available big data technologies are presented in detail – together with the identified extensions needed for Melodic. The requirement analysis and the technology evaluation together resulted in a preliminary architecture of the Melodic middleware platform. This preliminary architecture, as presented in this document, will lay the foundation of the first Melodic release, the integration release, due by the end of the first year of the project.
The Architecture definition details and explains the overall architecture of the MELODIC middleware, and in particular describe the key components of MELODIC, and how these components are to interact and collaborate in order to realise an efficient multi-cloud middleware platform for data-intensive applications. In addition, the initial feature definition of MELODIC is provided.
This deliverable provides scheme definitions, and detail related concepts regarding metadata management for both data stored in the cloud as well as user defined data annotations, including the security vocabulary needed to support appropriate access control mechanisms.
This report starts with a review of existing data management techniques for clouds, before reporting on the novel MELODIC methodologies for data placement, data migration, and post-migration handling of data in a distributed multi-cloud infrastructure.
Abstract
A report on the first iteration of the MELODIC editor and the metadata schema management for maintaining all the MELODIC design-time artefacts. The necessary GUIs will be developed and made available to administrator(s) for editing and managing all the MELODIC models.
An analysis of the MELODIC pre-migration, migration and post-migration phase methodologies to be considered for supporting the complete data and data-intensive application life-cycle management. It provides the appropriate business logic for supporting this life-cycle management in distributed cloud environments.
This deliverable focuses on the provider agnostic interface definition and mapping to the system requirement analysis of D2.1. Representing the first iteration this definition and mapping cycle a first prototype of the provider agnostic interface mapper is developed. This prototype builds the starting point for the resource management framework and it is also evaluated against the the first draft of the resource management framework.
This deliverable focuses on the first prototype of the resource management framework. It revisits the provider agnostic interface definition to provide an interface for the Upperware, including management and monitoring capabilities. This first iteration of the resource management framework also includes the completed integration of PaaSage and CACTOS components.
This deliverable finalises the resource management framework. It revisits the prototype of D3.3 and refines the resource management framework due to lessons lerned during the development of the data processing layer and the Upperware. The resource management framework is evaluated using of the data processing layer.
This deliverable describes the first iteration of the data processing layer resulting in a first prototype. Based on D4.1 and the parallel development of the rsource management framework it lifts a first data processing engine on top of the resource management framework and implements a first version of the data storage infrastructure. This first prototype is evaluated against the first iteration of use case implementation (D6.2).
This deliverable presents the final data processing layer of the Executionware. The data processing layer enhances the Executionware with concepts for enabling the native deployment and orchestration of Big Data processing frameworks, Function-as-a-Service (FaaS), container-based applications and data storage services. These concepts are implemented for the Big Data Processing framework Apache Spark, the AWS FaaS service Lambda, Azure Functions and for general Docker-based applications.
Authors:
Antonia Schwichtenberg, Katarzyna Materka, Kyriakos Kritikos, Paweł Skrzypek, Michał Semczuk
Abstract
The efficient and appropriate usage of existing opensource software in Melodic needs a careful study with respect to the changes and adaptations required for their integration with Melodic. This deliverable is aimed at identifying such changes. The frameworks reviewed in this deliverable include PaaSage and Cactus/Cloudiator. The review methodology is designed based on existing functional and non-functional features in each framework and the assessment of implementation code for each framework. The functional features are studied based on the objectives of Melodic: multi-cloud deployment of applications, cloud-agnostic description of applications and infrastructures, and support for data awareness and data locality, to name the most important ones. Performance, reliability, maintainability, and serviceability are among the important non-functional features considered in our methodology. The code assessment is performed by verifying the quality of the code against best programming practices as well as quality assurance requirements. We use the Clean Code approach as the best programming practice for evaluation.
Authors:
Antonia Schwichtenberg, Sébastien Kicin, Katarzyna Materka, Somnath Mazumdar, Jörg Domaschka, Yiannis Verginadis, Michał Semczuk, Paweł Skrzypek, Sebastian Schork
Abstract
A careful design of the Melodic integration strategy is very important as the main mission of Melodic is to integrate and adapt underlying frameworks, including PaaSage and CACTOS. This deliverable provides a detailed description of the requirements of this integration. The methodology for collecting integration requirements focuses on identifying the requirements separately for Control & Data Plane and Monitoring Plane. The other key set of requirements is functional testing and UI components testing, which specifies rules and conditions verifying the proper implementation of respective functional features of the Melodic system. The functional testing requirements are provided in the form of testing scenarios, which should be executed for a given Melodic release. These scenarios include initial deployment testing, global reconfiguration & local reconfiguration testing, metric management testing, reasoning related testing, and API testing. D2.1 “System specification” and use-case application descriptions are two main sources of identifying testing requirements. Apart from functional, also non-functional testing requirements are covered to verify the proper implementation of the nonfunctional features of the Melodic system. The nonfunctional testing scenarios include fault handling testing, performance testing, security testing, and other nonfunctional testing.
This deliverable’s released 1.0 and 1.5 of integrated MELODIC framework. It also provides a report of integration and technical testing activities (functionality, performance) based on test strategy and test plan included in D5.6 as well as the description of the initial version of testing infrastructure.
This deliverable is the second release of integrated MELODIC framework along with the security and DLMS technical components. It includes the feedback from testing releases 1.0 and 1.5. It also provides the report of integration and technical testing activites (regression, funcionality, performance and security tests) for release v.02 based on test strategy and plan included in D5.6.
Authors:
Małgorzata Jakubczyk, Marcin Prusiński
Abstract
This document presents the most important processes and tasks related to Quality Assurance of software developed in Melodic. In particular, the document should be used as a manual for performing all activities within testing processes. Also it should be used in conjunction with the ‘D5.06 Test Strategy and Environment’ deliverable, which contains the strategy of testing, acceptance criteria, test related products and so on. The deliverable’s intended audiences is as follows: It could be used by all participants in the project. Especially it shall be use by development teams (bugs handling process), test teams (bugs, test cases processes), architects (bugs, test cases processes), use case application users (test reporting, bugs handling) and management of the project (status of the testing of release). Altogether, the document is a complete and comprehensive manual for the most important software quality assurance tasks in the Melodic project.
Following the overall Melodic evaluation framework described in Deliverable 6.1, this Deliverable D6.2 describes the first implementations execution and feedback of the current release of the Melodic platform.
Following the overall Melodic evaluation framework described in Deliverable 6.1, this Deliverable D6.3 reports the second implementation and execution of the use cases and feedback of the release 2.0 of the Melodic platform.
The MELODIC platform isdeveloped as a multi-and cross-cloud middleware platform, enabling cloud application deployments in the most effective way from the point of view of the owner of the application.
In the scope of the MELODIC evaluation work package (WP6) and following to the data collection methodology presented deliverable D6.4, this document presents the results of the evaluation process based on the Melodic evaluation framework.
This deliverable reports all the dissemination activities taken up during the project. It includes actions taken up by academic partners (list and brief presentation of scientific publications, list of scientific conferences and EC workshops attended) and commercial partners (list of trade fairs and brief presentation of activities using different communication channels specified in D6.3).
This deliverable provides a draft plan for exploiting the MELODIC framework beyond the project. It includes both the individual exploitation strategies of the project partners and the joint strategy for the sustainability of the framework as a whole.
This deliverable provides an elaborated plan for exploiting the MELODIC framework beyond the project. It includes the specification of the lifecycle management methods & tools, sustainability model and licensing scheme as well as the individual exploitation strategies of the project partners.
Authors:
Geir Horn, Paweł Skrzypek, Feroz Zahid, Amir Taherkordi, Yiannis Verginadis, Katarzyna Materka, Michał Semczuk, Tomasz Przeździęk, Sebastian Schork, and Antonia Schwichtenber
Abstract:
This is a public report starting from the objectives, approach and challenges addressed by MELODIC, and documenting the results achieved, the lessons learned, and the recommendations for the future both with respect to technology, access and governance of the open source MELODIC platform, and promotes the use cases as examples for others to follow.
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 731664. 
