The main objectives of CLOUDSTARS are:

• Development and benchmarking of open-source technologies to advance the next generation cloud infrastructure, where it will play a vital role the development of new container technologies, as well as elastic and low-latency storage for ephemeral data, efficient resource management of hardware accelerators (GPUs, FPGAs and smartNICs) and networking.


• Design and benchmarking of novel cloud and edge serverless middleware to leverage the advances in the cloud infrastructure, which pivots upon all sorts of containerized technologies such as Function-as-a-Service (FaaS) platforms, serverless containers, and event-based orchestration to open the serverless execution model to HPC and analytics workloads. The project will also explore how the platform should protect and govern data throughout its lifecycle.


• Application of novel machine learning techniques (e.g., statistical and deep learning) for managing containerized cloud systems, involving the infrastructure and configuration of executions and services, and the optimization of data analytics platforms over serverless middleware and container infrastructures. The project will also explore the interplay between Edge and Cloud for enabling a diverse range of AI-enabled applications.