Executive Summary

This document provides a detailed update on Deliverable D1.1 "Detailed technical specifications for o2S2PARC, based on user requirements (e.g., from questionnaires) and technology evaluation" of the IT'IS SIM-CORE proposal.

The specifications for the SIM-CORE platform (o2S2PARC) were established through two separate activities: (i) contacting SPARC teams to establish their specific modeling needs (primarily by detailed interviews) and (ii) evaluating the best suited software technologies for a high-quality implementation of the SIM-CORE. The assessment of the software technologies was achieved by the comparison of various feasibility prototypes which were developed with different technologies. Deliverable D.1.1 and the related Milestones to this deliverable, M1.1, M1.2, and M1.3, have now been completed. As a result, high-level platform functionality specifications and concrete framework architecture specifications were derived that are summarized in this document.

Overall, the willingness of the SPARC teams to participate in interviews and to discuss team- specific needs and requirements was high. We have experienced large enthusiasm about both the possibility of sharing own models without the need to invest in developing user- interface functionality and the potential of getting access to powerful solvers, meta- modeling functionality, electrophysiology models, and analysis tools. Quality assurance aspects, data and model integration, user-friendliness, computational infrastructure, anatomical models, and support for workflows are other highly valued components. The platform concept, as originally developed throughout the SIM-CORE application process, is very well suited to meet these requirements, and only few extensions of the original proposal are required. These include: (i) support for "supercomputing", workflows, nerve microstructure modeling, thermal, biomechanical, and light propagation modeling (IT'IS does not have solvers for the latter two and would have to develop them), (ii) machine learning, (iii) additional animal anatomical models, as well as a detailed spinal cord model, (iv) modeling of fields generated by neural activity for sensing, and (v) further meta-modeling (control, model order reduction. At this point, other aspects do not seem to be of high priority for the SPARC teams (e.g., image-processing beyond segmentation and functionality for users to implement their own, advanced-interaction GUI-elements). Providing users with proper support will be fundamental to assure the widespread adoption of the platform.

As mentioned above, all of the requirements are very much aligned with our prior expectations and only a minor adaptation of the proposal is required. We propose, however, an additional sub-project to support the SPARC teams in adopting the SIM-CORE platform. We would further consider to add some of the proposed extensions (see above) after detailed discussions with the SPARC office. Furthermore, we propose to implement Deliverable 6.4 (development of functionality to allow users to create UIs for their services) with increased priority - to allow faster SPARC-team model integration - while delaying the implementation of high-level Python scripting interfaces (D6.1/6.2), user-defined viewer plug-ins (D5.4), and advanced image-based nerve tracing (D3.3).

Our evaluation of existing software technologies resulted in clear favorites for the technologies and approaches to be used for implementing the SIM-CORE platform. These are: (i) a front-end built with [qooxdoo], (ii) a python-based web-server, and (iii) a swarm of pipelined computational services deployed inside dockers and (iv) orchestrated by a director that is in turn interacting with the web server. This design provides the SIM-CORE with sustainable capabilities that enable scientists to share their computational models in dedicated environments with minimal adaptation effort while providing at the same time the capability to build workflows with other scientist's models and an attractive interface to established models/services.

The planned design of the platform will enable maximal flexibility with regard to the wide variety of existing and envisioned user-generated modeling services. It will also offer user-friendly interfaces, allowing users to engage on different levels with the platform, ranging from simple execution of existing models to the advanced generation of services with fine-grained control over dedicated user-interface elements, depending on expertise. Due to the modularity of the chosen approach, it will be simple to extend and adapt the platform at later time-points, and it will be feasible to revisit some of the choices as technology evolves and to replace layers or components of the implementation without the need for a complete redesign of the platform.

Introduction

This document reports on Deliverable D1.1, "Detailed technical specifications for o2S2PARC, based on user requirements (e.g., from questionnaires) and technology evaluation", of the IT'IS SIM-CORE proposal. Deliverable D.1.1 and the related Milestones M1.1, M1.2, and M1.3 have been completed. After presenting the requirement gathering approach and effort, the obtained information is summarized, and conclusions are drawn. The technology evaluation approach and activities are introduced, results are presented, and conclusions are reached. From this, high level platform functionality specifications and framework architecture specifications are derived.