UNICORE Summit 2018

Scope and main topics

The UNICORE Summit is the annual meeting of the UNICORE community. It provides a unique opportunity for UNICORE users, developers, administrators, researchers, service providers, and managers to meet.

Participate to share your experience, present recent and planned developments, learn about the latest UNICORE features, and get new ideas for interesting and prosperous collaborations.

The areas of interest of the UNICORE Summit 2018 were:

applications and workflows
high-performance computing and big data processing
security, identity management, federations
portals, science gateways, interoperability
cloud and virtualization techniques
operations and deployment of UNICORE
recent and planned development work
new ideas and concepts

Schedule and Presentation Slides

The meeting took place as video conference on September 21, 2018 from 10:00 to 12:00.

Friday, September 21, 2018
10:00 - 10:10	Welcome
10:10 - 10:40	Timo Strunk (Nanomatch) Removing complexity layers in the UNICORE installation	Presentation
10:40 - 11:10	Björn Hagemeier et. al. (Jülich) UNICORE Puppet Module	Presentation
11:10 - 11:40	Bernd Schuller (Jülich) UNICORE 8 development direction	Presentation
11:40 - 12:00	Open discussion
12:00	Wrap-up

For further information please send e-mail to unicore-summit[at]unicore[.]eu (unicore-summit[at]unicore[.]eu)

A Look Back

The UNICORE Summit 2018 took place on Friday, September 21, 2018.

For the first time this year’s UNICORE Summit was organised as a videoconference. 14 people participated in the meeting contributing to presentations and lively discussions.

The ease of installation and configuration of the UNICORE services was addressed in the first two talks. Firstly, Dr. Timo Strunk from Nanomatch, presented a new installer for UNICORE. It offers quick installation and secure configuration of UNICORE and Unity within 20 minutes and is available under BSD licence on GitHub (https://github.com/timostrunk/UNICOREDaemonCerts). Subsequently, Björn Hagemeier from Forschungszentrum Jülich reported on experiences of using Puppet for managing UNICORE installations. This approach is currently used on supercomputers in Jülich and automates the installation and configuration of UNICORE services to a high extent. The Puppet modules are going to be published as soon as the documentation is finished.

Dr. Bernd Schuller from Forschungszentrum Jülich discussed the development directions for the next major release of UNICORE. The focus will be on strengthening the core use cases, such as support for heterogeneous clusters, and making the REST API the primary interface. The complexity of installation will be reduced by removing unused features and simplifying configuration files.

During the discussion the need for simplifying installation and configuration has been stressed repeatedly. Valuable user feedback also led to identification of most urged features that should be implemented. Examples are a mechanism for restarting workflows and a job model which is compliant with containers and thus better suited for cloud environments. Furthermore, migrating the source code management system from subversion to git has been suggested and will be considered by the core development team.