— WP objectives
Leader: Philippe Merle
- Development of exact methods for verifying and optimizing cloud infrastructure and application models
- Introduce the notions of priority and temporality to model the variability of application models and cloud infrastructures
- Specify modeling mechanisms to take account of application/system variability
- Definition of an institution for interaction models, including composition operators and a framework for compositional refinement
- Modeling of capability-based Cloud service “bouquets” for supplier lock-in discovery and resolution
- Development of an integrative IDM framework
— Missions
This work package addresses a number of issues related to the complexity induced in Cloud-Edge application and infrastructure models: formal verification and optimization of these models, multi-layer variability, the relationship between model expressiveness and efficient solution computation, lock-ins of proprietary models and heterogeneity of Cloud-Edge application and infrastructure modeling languages. The project aims to develop accurate methods for verifying and optimizing cloud application and infrastructure models, and to introduce notions of priority and temporality to model the variability of cloud application and infrastructure models, to specify modeling mechanisms to take account of application/system variability, to define interaction models including composition operators and a framework for compositionality-aware refinement, and to model capacity-based Cloud service “bouquets” for the discovery and resolution of supplier lock-ins. Finally, the work package aims to develop an integrative IDM framework/workshop.
— Tasks description
T1.1 Logical reasoning for cloud infrastructure and application models
This task will provide models of cloud applications and infrastructures via formal models, study the impact of theoretical advances in logic on the management of cloud applications and infrastructures expressed by formal models, verify properties such as the consistency of these models, and integrate various optimisation issues. The problem addressed by this task is that modelling languages for cloud applications and infrastructures are often highly expressive, making the verification of properties and the optimisation of these systems described by symbolic constraints complex or even undecidable. Given recent theoretical developments on the decidability of certain logic fragments, it now seems possible to obtain exact methods for certain constraints in models describing applications and cloud infrastructures. We propose to approach optimisation problems through the knowledge compilation approach and to associate them with optimisation practices around control.
T1.2 Managing variability in cloud infrastructure and application models
This task aims to take better account of the variability of cloud systems, mainly on two levels.
- On the one hand, the deployment of (micro)services in the cloud relies on a set of layers, each with its own configuration constraints. The global configuration is therefore the result of a compromise of local configurations (i.e. specific to each layer), where the variability of several interdependent systems has to be managed simultaneously. This task will therefore study the notions of priority and temporality applied to variability to enable the designer and developer to define different configuration strategies (evolutionary, permissive, etc.).
- On the other hand, this task will also aim at reconciling variability and placement, which are usually treated independently. Application description models should allow the choice of the placement algorithm to be used, which may change depending on the resources, the target architecture (cloud vs. fog) or the size of the infrastructure (centralised vs. distributed solution). The goal is to design application description models that take variability into account, while at the same time enabling the design of efficient placement algorithms.
T1.3 Communications modeling and network typologies for cloud applications and infrastructures
Christophe Gaston
Cloud computing is gaining popularity, including the deployment of mission-critical systems. As a result, it is important to have a formal framework for reasoning about distributed systems composed of subsystems deployed on different machines and interacting through message passing to realize services. We propose to extend recent work that associates an “integration language” with several formal semantics and proves their equivalence, by defining a formal specification framework based on interactions and dedicated to the specification of systems deployed in clouds.
Furthermore, an important aspect to consider in applications from a consumer point of view, in the interest of good practice, is to limit the ability of suppliers to impose a lock-in policy. In this task, we propose a model that allows the injection of rules to identify all the constraints associated with a required resource and make them observable to users right from the discovery phase; this can anticipate a priori not only the detection of configuration problems, but also of lock-ins.
T1.4 IDM framework for modeling cloud applications and infrastructures
Chokri Mraidha
The rigorous management of cloud applications and infrastructures requires the use of formal modeling languages adapted to each phase of the lifecycle and to each property of interest to be verified. The goal of this task is to establish an integrating framework of model-based approaches for the design and management of cloud applications and infrastructures by all project partners. This framework will have to ensure interoperability / integration of the different specific languages (DSLs) developed / used to cover the engineering and management needs of cloud applications and infrastructures. This will require the implementation of a suitable meta model and appropriate model transformations.
Other Workpackages
