This deliverable presents how RECAP use cases are being implemented in the RECAP simulation framework, that encompasses of two different simulation approaches in order to cover all requirements defined previously in D3.1 Initial Requirements: Discrete Time Simulation (DTS) and Discrete Event Simulation (DES). After analysing all uses cases, and considering the advantages and disadvantages of each simulation approach, it was decided that large-scale scenarios will be modelled by using DTS, while scenarios with more peculiar details will be modelled by using the DES approach, as follows:
Discrete time simulation (DTS)
- Use Case C: Edge/Fog Computing for Smart Cities (SATEC)
- Use Case D.1: Network Function Virtualization – Virtual Content Distribution Networks (BT)
Discrete event simulation (DES)
- Use Case A: Infrastructure and Network Management (TIETO)
- Use Case B: Big Data Analytics Engine (Linknovate)
- Use Case D.2: Network Function Virtualization – Cloud Connect Intelligence (BT)
As intermediate results, this deliverable presents the Use Case D.1: Network Function Virtualization – Virtual Content Distribution Networks (BT) modelling and results. This use case is based on the BT network topology, that is composed of 1132 sites, distributed in a hierarchical 4-level architecture. From experiments, we show that the DTS simulator is able to run large-scale scenarios such as the vCDN use case in a timely manner.
The simulation application models and logical simulation experiment assumptions are presented for the rest of use cases. The application models are implemented and demonstrated by using a unique RECAP produced application and workload meta models which allow distributed fog application behaviour description within simulation framework.
To validate the accuracy of RECAP simulation modelling techniques the real system data was compared with simulation output. In the first phase of validation exercise, controlled NFV benchmark experiments were executed by Tieto inside their testbed by applying emulated user workload to VNF chain. After the monitoring data was gathered from the experiment, the second validation phase was executed by modelling and executing simulation experiment using the same system configuration. Finally, measured and simulated results were compared side by side.
The integration with the optimiser framework is also part of this deliverable. The RECAP simulation framework has the capability to receive (a set of) JSON files containing a simulation configuration, execute it and provide the required output. Experiments were carried out considering the vCDN use case.
The next steps for the RECAP simulation framework focus on modelling other remaining use cases taking into consideration the requirements defined in D3.1 Initial Requirements, and complete the integration with optimiser framework by using a well-defined API.
To download this deliverable, please fill in the form.