Project Overview
The PHYSICS project will provide:
- Graphical view of the application flow with drag and drop commands to facilitate the creation of pipeline of functions (FaaS)
- Abstraction and application component reuse through intuitive flow programming approaches incorporating reusable design patterns
- Global Continuum Layer to easily deploy functions, organizing multiple application objectives at the same time in a multi-cloud and hybrid cloud infrastructures
Who are we?
PHYSICS is a joint effort of global leaders in ICT integrators, technology producers, research centers, and universities with a mutual aim to showcase the potential of the Function-as-a-Service paradigm and foster its wide deployment by application developers, platform providers and cloud service providers.
More information on PHYSICS’ consortium is available here
PHYSICS’ value proposition
PHYSICS aims to unlock the potential of the Function-as-a-Service paradigm through a vertical solution that will enable:
PHYSICS’ motivation
Cloud Computing Challenges:
What is FaaS?
It is a serverless way to run applications, where you do not have to worry about defining the running infrastructure and where you only pay when functions run. Functions are defined in workflows and are activated through an event manager.
Pilots
PHYSICS pilots have been carefully selected to cover the three key spheres of European daily life and economic activity:
| Domain | Smart Manufacturing | Smart Agriculture | eHealth |
|---|---|---|---|
| Scenario | i) Deployment of substitute services in the cloud ii) High Confidence Quality Control | i) Data Collection and deployment for new greenhouses ii) Performance & Simulation | i) Inference using predictive models ii) Phenotyping using a set of generative models iii) Data synthesis using a set of generative models |
| Deployment | Edge Deployment for low latency Cloud Deployment for redundancy and on demand computation power | Edge Deployment for Data Collection Cloud Deployment for Simulation | Cloud |
| Focus Objective | Reliability of the service | Reliability Computation Time | Near real time response regaless of load for the ML algorithm |