D6.1. Initial Workload, Load Propagation and Application Models

RECAP's work package 6 (WP6) is dedicated to application modelling, workload modelling, and workload propagation modelling. This first deliverable of WP6 aims to describe an initial version of these models. The achievements to date of the WP are presented throughout the document with the following outline:

1. An analysis of issues and challenges of the modelling tasks;
2. A detailed description of application models together with mathematical models for each system/application introduced in the use cases;
3. A detailed description of workload models for each use case;
4. A sketch of how to use the models for system simulation and optimisation.

The key parts of the document reveal application and workload models of the system defined by each use case in different aspects. Particularly, a system architecture shows the system's implementation and its operational model, and also visualises the traffic flows in the system. An application model exposes the behaviour of the system and also provides means to estimate various KPI metrics depending on the nature of each system and its business requirements. Workload models demonstrate how the load looks like over time and are also used for future load predictions.

The document also outlines the utilisation of the models in other WPs to implement the RECAP simulation framework (WP7) and to address optimisation problems (WP8). More specifically, WP7 uses the models as a base to simulate how the systems work and how they are affected by the variations of the workload. WP8 uses the models to define optimisation cost functions, utility functions, and constraints of the optimisation problems and then devises solutions to improve the system performance according to the business requirements. Finally, the document sketches the applications of the models together with system optimisation solutions to realise the system elasticity and remediation. Additionally, the need for further fine-grained models is also mentioned here. All of these efforts are ongoing and will be advanced in the forthcoming stages of RECAP.