PROCESS ARCHITECTURE

Processing of exascale data faces common problems: extreme data sizes and data structure complexity.

-The use cases have confirmed the increasing requirements in terms of data storage and processing. At the same time, final users are demanding for adaptable and easy-to-use solutions.

-An architecture to tackle these demands and draw exascale data near users (SMES, researchers) is key for the success of the PROCESS project.

-The solution must be able to decouple access to the data from their use in computation and to provide simplified access to individual components of the data

APPROACH

- Scalability and modularity for high distribution and parallelism.

- Sub-modules to exploit heterogeneus resources for versatility.

- Containerisation, instead of virtual machines.

Reference exascale architecture

Selected technologies

MAIN FEATURES

- Modular software platform capable to handle exascale datasets

- Microservice infrastructure: customazible, based on independent elements

- Interaction through RESTfull APIs (mainly)

- Authentication by security tokens

STRUCTURE

- Virtualization layer

- Data services

- Computing services

REFERENCES



DELIVERABLES

- D8.4 Final release (pending)

- D8.3 First packaged version of the PROCESS software

 

PROTOTYPE

 – (Pending)