University Of Gothenburg’s Use Of HR Open Standards For Provisioning

Guest Blog Post: Ola Lundgren is an Integration Architect | University of Gothenburg

Founded in the late 19th century, and currently home to upwards of 50,000 students every year, the University of Gothenburg is one of the larger universities in Sweden. The University has an employed staff of roughly 6,500 teachers, researchers, and administrative personnel, as well as thousands of research students and external staff linked to the University in various capacities. As a result, the challenge of keeping a consistent view of identities, and their various affiliations to different parts of a sprawling organization, across a myriad of different IT systems grows larger every year.

From an IT architecture perspective, our integration landscape is heavily focused on service orientation and re-usability. This has a number of implications for the way we design integration solutions, and a key element is the application of standards. For our core provisioning services, we try to use standard message formats­ for the particular domain. We have found that this shortens development time when implementing new integrations and minimizes misunderstanding within the use of common concepts and nomenclature.

In the spring of 2019, we launched a project with the purpose of implementing a new system of record for the domain most adequately described as Person/Organization. An expressed purpose of the project was to provide a platform of integration services for aggregated organizational affiliations, whether they be employed or external staff, researchers, teachers or students of various kinds – in many cases several at the same time! Keeping with our principles, we started to look around for available standard formats to base these services on.

Our solution was to take our existing information model for this domain, along with a list of basic requirements, and search for alternatives. The initial list of candidates contained existing local canonical models, data formats published by Swedish national agencies, and global standards. Pretty quickly, however, we found that most formats were either too thin (i.e. lacked coverage in a lot of areas of our information model which would be a major extension driver), too verbose, or in other ways were not suited to our use case.

What we ended up with was the HR Open Standards Common library for Person, Affiliation, and Organization.

Among our basic requirements were:

• Structured format: based on XML or JSON
• Possible to validate messages against schema
• Open standard: vendor independent
• Extensibility: possible to make local additions
• Localization: support for multi-language text fields

Not only did the HR Open Common types check all of these boxes, they also provided the necessary information coverage for “out of the box” data we needed to represent (Person, Affiliation, Communication, Security Credentials, etc.). Being an academic organization with a lot of research staff, we appreciated the built-in support for academic publications and educational merits, as well as the how easily the standards were extendable with both simple types and complex structures.

Adding to our support for the HR Open Common types, was its flexibility in the right areas. Other models were often too rigid or geared to a narrower use case, for example explicitly defining entities like Employee and Employment, or only providing one organizational affiliation for an individual. Since our extended use case involves a generic platform for organizational affiliation, the relevant individuals could be any type of staff, employees, or non-employees, guest researchers, or students, and even sometimes several of these choices at once!

In addition to providing an elegant way of describing this relationship (a Person with any number of Affiliations to Organizational entities), the HR Open schemas proved to be technically sound with the right amount of verbosity for a standard. The toolbox includes great extras like generic identifier types, structures for different types of contact information, language enabled strings, and so much more that we can continue to use as extensions in our messages.

We are very happy with this design choice for our integration services and look forward to seeing how the standard continues to develop in the future.

ABOUT THE AUTHOR

Ola Lundgren is an Integration Architect at the University of Gothenburg and has worked there since 2015. He has a M.Sc. Computer Science & Engineering with experience in Software Engineering and IT architecture in the Education sector.