E4 Computer Engineering and their role in DECICE project
E4 will contribute to DECICE with their gained experience in the domains of storage, data centre/HPC, use cases, cloud infrastructures and networking.
E4’s role in this project is to:
- integrate testing task and
- lead the activities on deployment, validation, and performance assessment
E4 Computer Engineering designs and manufactures highly technological solutions for HPC clusters, cloud, data analytics, artificial intelligence, and hyper-converged infrastructure for the academic and industrial markets. The company has been collaborating for years with the main research centres at national and international level (CINECA, CERN, ECMWF, LEONARDO) and is involved in national and European projects in the HPC and AI fields (EuroHPC JU, EPI, EUPEX, Horizon Europe, KDT).
Through sister company E4 Analytics, E4 works to integrate Data Science in organizations undertaking the Digital transformation of their business to improve products/processes and optimize resources. It operates at the intersection of business and technology, supporting the customer in the adoption of AI solutions: with E4 Analytics, the data business become a resource to produce value.
Each E4 solution is unique, tested in every single component; validated to verify the actual performance of each system and serviced by technicians who provide assistance in the most extensive and complex Italian and European computing infrastructures.
E4 contribution to the DECICE project
E4’s role in DECICE project is to integrate testing task and lead the activities on deployment, validation, and performance assessment. To do this, E4 will use the experience gained over the years in the domains of storage, data centre / HPC, use cases, cloud infrastructures and networking.
E4 will lead the Work Package on “Deployment, validation and performance assessment” by focusing on the following tasks:
• “Definition of test applications and use case requirements”, in which a list of use cases and test applications from different stakeholders will be compiled and used throughout the project to define the requirements for the test bed, test the environment, evaluate the performance, and provide feedback to the developments.
• “Project development environment” will design a development, testing and validation environment where prototypes and models can be deployed and tested, with a full and manageable access to hardware, software, and cloud infrastructure.
• “Deployment”, which includes the actual deployment of the project development environment and the application of use cases. This action will take place in three stages:
a. Preliminary deployment: in this phase an infrastructure will be deployed, and a preliminary but representative subset of use cases will be run on the system to stress test the entire environment and highlight any inconsistencies or shortcomings with respect to the target performance and functionalities.
b. Testing: in this phase the infrastructure deployed in the previous stage will be connected with a similar infrastructure in a different location to simulate dynamic access to an interconnection with distributed resources across partners.
c. Full deployment: the project development environment will be tested in its final configuration with geographically and functionally distributed resources including public clouds in Europe and the infrastructures provided by the partners. All project partners will have access to the project development environment.
• “Performance evaluation”, where an in-depth operational evaluation of the DECICE framework for the specified use cases will be implemented and key performance metrics will be collected and analysed to evaluate the performance of the framework and guide the development. The evaluation will cover regular and uncommon system states to detect performance bottlenecks. The results of these evaluations from all three phases will be used for the continuous refinement of the framework and its integrated components.
Besides, E4 will also contribute to the “Open framework and virtual training environment”, leading the integration testing task, which will design an integrated testing environment focused on the replication of the architecture using mock elements to test the functionality of the individual components.
Furthermore, support will be given to “Cloud management framework integration”, by focusing in particular on the integration testing, building a Continuous Integration and Delivery pipeline for consistent builds of the DECICE API and the integration into Kubernetes/KubeEdge. Therefore, integration tests for the DECICE components will be developed and good software development practices will be enforced.
E4Company, E4ComputerEngineering, DECICE, EuroHPC, HPC, Cloud