Design environment for extreme-scale big data analytics on heterogeneous platforms (EVEREST)
H2020 Research Project to develop an accelerated programming environment for extreme-scale big data applications
Big data and Artificial Intelligence (AI) driven applications such as data analytics, weather/traffic forecast, etc. demand a huge amount of computational power. In this scenario, a hardware accelerated heterogeneous and distributed architecture that combine Cloud, Edge and IoT computing is key to collect and process an enormous amount of data.
The EVEREST H2020 research project aims at addressing this challenge by providing a virtualized programming environment for extreme-scale applications. More in particular, EVEREST project ambitions to optimize big data and AI applications data management and processing through:
A data-driven design and programming framework for software generation. Domain-specific languages (DSLs) are used by application designers to express the logic of the algorithm at a higher level.
A comprehensive, distributed and virtualized runtime environment composed by multiple heterogeneous hardware accelerated nodes. The EVEREST virtualization environment can configure the underlying hardware to execute alternative code variants generated during the compilation phase.
This virtualized runtime environment represents the Virtual Open Systems focus activity in the project. In fact, the company designs and develops virtualization extensions for both ARMv8, RISC-V and x86 CPU architectures. Such extensions will expose hardware configurable parameters directly to the applications from within virtual machines as well as they will support heterogeneous multi-node environments, enabling a seamless integration of both servers (typically Intel x86 machines) and Edge nodes (low power ARMv8 and RISC-V devices).