Development of a Semantic IT Service Catalog in a German Public Organization

Co-presented by Jonas Jetschni B. Sc., University of Applied Sciences Brandenburg

IT operations and management in German public organizations – as well as in small and medium enterprises – are mostly homegrown, heterogeneous, and primarily focused on technical aspects. Their performance can be characterized as rather reactive than proactive. Without clear central IT organization and governance there is a lack of transparancy and guidance for users as well as for staff and for business executives. In this regard, universities – particularly small ones – are very typical public organizations. The concept of service orientation and the implementation of sustainable IT service management can be considered as state-of-the-art approach for solving such kind of problems.

One of the recommended first steps towards service orientation is the implementation of an IT service catalog, which contains structured and user-group-specific information about all relevant IT services at the organization. Such a catalog not only displays all provided for different users IT services together with their characteristics and options, moreover, it allows substantious judgement about a lot of questions, relevant for the management of the organization, e. g.:

• Are there redundand systems and services in the organization which shall be streamlined?
• How shall be set the different IT cost types against a definite service, provided by one department for the benefit of the over-all organization?
• Which qualifications of IT staff or which guidelines and policies are missing for providing IT services at definite (agreed) service levels?
• Can the organization benefit economically from standardization or outsourcing of IT systems and services? 

A non-representative analysis of service catalog implementations in german public organizations reveals the situation as follows. A big part of them don’t have a service catalog implementation at all. Most existing implementations are traditional Content Management Systems (CMS) like Typo 3 or Drupal. Some use plain documents for the listing of IT Services, mostly without group-specific detailed information. Very few apply rather technical solutions, based on Configuration Management Databases (CMDB).

Based on litarature analysis and expert interviews the following performance and maintenance requirements for an IT service catalog implementation in a public organization was collected:
1. Flexible adaptability over time
2. Moderate effort for implementation and maintenance
3. Adaptability with respect to different user groups
4. Easy integration in the organizational information and knowledge management
5. Visualization of relations between different IT systems and services
6. Broad reuse of relevant organizational data

None of the three mentioned implementation types satisfies all of the listed requirements. A document-based system lacks mostly in adaptability and integration features. CMS lacks primarily in visualization of relations and in reuse, whereas CMDB-based implementations are intended for technical audience and not for non-technical users. That’s why we came to the decision, to develop and implement a semantic IT service catalog. 
The first step of that development is now finished. A vertical prototype was developed, implemented and tested as well as formally evaluated by members of different user-groups. The GUI provides three user groups (datacenter experts, teaching staff and students) with access to IT service information along their main competence questions. The application level is build upon a thoroughly developed, standard-based, lean and at the same time semantically rich, reusable ontology. Although not all relevant concepts and releations could be gathered from standard or common vocabularies, a broadly mapping to standard concepts and relations was realized. On the data level a partial implementation of relevant data from different sources (database extraction, document parsing and direct insertion) was realized. 
In the course of development different approaches and tools was discussed and tested. For ontology engineering was tested seven open source as well as proprietary tools, four of them (CMap, Protégé, rdfEditor, OWLgred) where used for different stages and requirements in the engineering process. For application development some basic tools and approaches where chosen from Business Ontology Toolkit (http://ontology.it/tools/botk4/) completed by a set of specific servers, like D2RQ and Fuseki.

Further development is mainly oriented in enhancing the data base and implementing more analytical features for mentioned user groups. Later on the semantic IT catalog shall be extended to other user groups, like executives, prospects, business partners, and administrative personnel. Another line of further development focus on cooperation with other public and healthcare organizations.