ACCORDION and CHARITY projects, funded by HaDEA under #2020, will share innovative insights about "Cloud-Edge Continuum Resource Management and Application Steering" which are common elements to both projects.

Don't miss this opportunity!

Computing Continuum management is a fast growing field of techniques and tools,  aiming at the joint exploitation of Cloud and Edge infrastructures together to  fulfill requirements from a variety of applications. In this seminar we propose the different takes at Continuum exploitation of two  H2020 projects, ACCORDION and CHARITY.  Both projects aim at leveraging  information about the application structure and its needs (in terms of  resources, constraints, QoS targets, and QoE ones) to dynamically match the  available resources at all levels of the Continuum platform. Information  continuously gathered from the platform allows to plan and steer distributed  applications and their allocation tradeoffs in near real time to satisfy the  user expectations. Edge exploitation for services is a powerful tool to spread  computation load and reduce service latency, but requires utter care in enabling  local acceleration and managing more constrained resources. We will introduce the viewpoint, main goals, design choices and current results  of both projects. ACCORDION looks for a generic support to multiple application  structures and application fields, where the formalisation of the application  constraints and goals is essential to exploit the information dynamically made  available from all parts of the platform.  By comparison, CHARITY looks at the  needs of AR, VR and Holographic applications. The problem viewpoint thus shifts  toward a class of applications that place quite daunting constraints on the  allocated resources, mandating the use of more sophisticated and targeted  allocation and co-scheduling techniques to closely deal with complex network,  storage and computing requirements, while also satisfying application-specific  needs like Edge-based 3D rendering acceleration and caching.