Latest Releases

A European Federation Software Suite

UNICORE (Uniform Interface to Computing Resources) offers a ready-to-run system including client and server software. UNICORE makes distributed computing and data resources available in a seamless and secure way in intranets and the internet.

UNICORE has special characteristics that make it unique among middleware systems. The UNICORE design is based on several guiding principles, that serve as key objectives for further enhancements.

Latest news

14 Dec 2022

UNICORE 9.1.0 released

Version 9.1.0 of the UNICORE Core Servers and the UNICORE Commandline Client are available for download from SourceForge.

 

What’s new

  • NEW FEATURE: port forwarding / tunneling (read more)
  • IMPORTANT BUG FIX: some jobs were already launched without waiting for files from the client
    (despite job having ‘haveClientStageIn’: true) due to a JSON parser bug
  • Improvement: take requested login node into account also for pre/post processing
  • Updated third-party libs

 

17 Oct 2022

UCC 9.0.0 is available

The latest version 9.0.0 of the UNICORE Commandline Client is available for download from https://sourceforge.net/projects/unicore/files/Clients/Commandline%20Client/9.0.0.

Apart from support for the new features in UNICORE 9.0, such as (single-site) workflow support in every UNICORE/X and Slurm allocations, this release offers cleaner command-line options and a smaller download size. The documentation can be found on unicore-docs.readthedocs.io.

Note that this release requires Java 11 or later!

30 Sep 2022

UNICORE Core servers 9.0.0 released

Version 9.0.0 of the UNICORE Core Servers is available for  download from SourceForge https://sourceforge.net/projects/unicore/files/Servers/Core/9.0.0.

We used this major release to remove a lot of “baggage”, update many of our third party dependencies and make UNICORE simpler to configure and manage. In terms of new features, every UNICORE/X server now comes with workflow processing capability for easily running single-site workflows. Furthermore, support for “interactive computing” via Slurm allocations has been added. Last not least, directly integrating UNICORE with a Keycloak OIDC server for user authentication is now possible.

Visit our new documentation pages at https://unicore-docs.readthedocs.io

What’s new

  • This release requires Java 11
  • Workflow processing feature in UNICORE/X (single site workflows)
  • Batch system allocation job support (Slurm only, via ‘salloc’)
  • Direct support for Keycloak OAuth server for user authentication
  • Support PostgreSQL as persistence back-end
  • Simplified and harmonized config files
  • Remove SOAP/XML endpoints
  • Bug fixes

 

Archive
  • Jupyter@JSC

    Interactive supercomputing based on UNICORE

10 reasons to use UNICORE

UNICORE comes with a history of more than 25 years. Originally initiated in the Supercomputing domain, today UNICORE is a general-purpose federation software suite. Based on open standards and technologies such as HTTPS, REST and JSON, it offers a rich set of features to its users. It allows resource providers a secure way of integrating HPC compute and data resources into new applications, solving the associated problems such as user authentication, user mapping and access control.

  • Mature, well-proven federation software suite from Europe
  • Integrated solution for federations of HPC systems as well as Clusters and storage systems
  • Fast and competent support directly from the developers
  • Easy to install and configure clients and services
  • Powerful command line client and Python API bindings
  • Application integration mechanisms on the client, services and resource level
  • Mature workflow support
  • High throughput data transfer
  • Source code is available and easy to modify
  • Easily extensible
UNICORE is widely used in European and international projects, initiatives, and infrastructures such as the Human Brain Project (HBP), the US Extreme Science and Engineering Discovery Environment (XSEDE) and the Polish Grid Infrastructure (PL-Grid). Companies use UNICORE for automation and speed-up of modeling tasks, e.g. Nanomatch GmbH.
UNICORE offers an abstraction layer for compute and storage resources. It can easily be configured to interface a variety of systems from high-end supercomputers to commodity clusters or single workstations. The same holds for storage systems from standard POSIX file systems to Amazon S3 or OpenStack Swift.
Our committed team of developers guarantees for stable software. We are closely watching all support channels such as mailing lists and bug trackers to provide timely answers and fixes.
All components of UNICORE, clients and services, are available as installation packages or archive files. The configuration files are clearly structured and come with sensible defaults. A test installation for quick evaluation can be up and running in minutes.
Seamless access to the compute and storage resources is provided by an command line client, that lets the user run tasks from simple jobs to sophisticated workflows, including the necessary input and output data, and easily monitor job execution and output retrieval. Programmers and application developers can use our Python library 'pyunicore' that makes using the RESTful APIs even easier, and which is conventiently available from the Python Package index (PyPI).
UNICORE is capable of supporting applications installed on the target systems. A description of the application with execution path on the target system, important parameter settings, and call parameters can be configured in the UNICORE execution service. On the client side such an applications can seamlessly be used in tasks by just specifying the necessary parameters.
The sophisticated UNICORE workflow support allows users to model workflows and manage input and output data, supported by powerful control structures.
The UNICORE file transfer (UFTP) service uses parallel input/output streams and dynamic firewall port opening based on FTP-compliant protocols. It offers optional data encryption, data compression, partial reads/writes to a file, and much more.
UNICORE is completely written in Java and Python and is open source. The sources are available on GitHub.
Developers are welcome to join the UNICORE developer community and contribute their ideas.

UNICORE is used in e-infrastructures of any nature and without limitations on the type of computing resource. Single PCs, cluster systems to leadership HPC systems like the once forming the PRACE Research Infrastructure are supported as well as a variety of storage types. Data management functionality includes high-performance file transfer and metadata handling. UNICORE‘s flexible security architecture enables many usage scenarios.

© Forschungszentrum Jülich 2023