System Installation and Setup
Initial RECAP setup using the RECAP CLI
This video shows the use of the RECAP CLI on an OpenStack infrastructure. The setup process of a RECAP installation is shown. On the left side the network topology in OpenStack, the right upper side displays the Rancher interface, while the right lower side shows the installation progress via the command line.
Deployment of RECAP monitoring and landscaping components
Assuming a RECAP installation from the first demo scenario, this demo shows the setup of RECAP components for monitoring the instances and the landscape of the RECAP installation.
The left side shows the Rancher interface, and further on the Grafana dashboard as well as the neo4j Landscaper displaying the infrastracture. The right lower side shows the setup progress via the command line.
AutoML Library Demonstration
jupyter-lab and scikit-learn Intergration
This demonstration shows the integration and use of the fastautoml library with scikit-learn and jupyter-lab.
Specifically, this shows data loading and calcuation of an optimal model for the Linknovate dataset.
Neural Networks Architecture
This demonstration shows training progress of the fastautoml library fastautoml library for a classification task on the MNIST dataset. The upper part shows the network architecture, while the lower part displays the training and validation scores.
RECAP Modelling and Autoscaling
This is a two-part demonstration that provides an overview of the modelling and application optimization capabilities of the RECAP runtime system. The demo showcases a data-driven workflow for construction of a city-scale model set integrating geographical information (street maps and structures), population census data, network infrastructure models (physical and cellular networks); as well as application (structure and deployment) data to provide a mathematical foundation for simulation-based experimentation and optimization of systems.
Dockerized OpenStack Demonstration
This video demonstrates the setup process of an OpenStack in a Dockerized environment in action.
Synthetic Data Generation
Generative Adversarial Networks
This video shows how Generative Adversarial Networks (GANs) are able to learn the underlying original data fidelity throughout several epochs.
RECAP has published an extensive dataset of artificial and real workloads.
Smart City Use Case Demonstration
This demonstration presents A fog computing and large scale IoT scenario for supporting Smart Cities.
- Models to distribute data and computational resources in the proximity of end users.
- Improvements in system management, efficiency and scalability over a fog computing architecture.
- IT automation for deployments.
- Support for large scale IoT scenarios.
- Edge/Fog/Cloud integrated architectures.
- Complete IoT Platform design.
- Trustworthiness of IoT systems.
Optimised Placement & Infrastructure Planning
The demo outlines the programmatic and analytical process for infrastructure optimisation i.e. short-term: mapping VNF service placements in the network under different scenarios to physical resources, and/or longer-term: requirement for and placement of physical assets in a network to meet predicted demand, in other words capacity planning.