In several national and even continental data strategies worldwide, the decentralized data space concept aims to address concerns of data sovereignty among organizations. A significant number of projects and a few already operational data spaces address application domains in industrial domains such as manufacturing, mobility and logistics, or health. However, data sovereignty has also become a key concern of artists and cultural institutions who are pursuing the two-pronged and sometimes conflicting goals of creating added value from data sharing, and protecting their intellectual property and personal privacy rights. Additional challenges of this sector include an orders-of-magnitude larger number of potential players compared to existing data spaces, frequently limited IT capabilities, a complex diferentiated system of data types and regulations, and a novel interplay between heterogeneous data integration and analytics with human creativity, among many others. Also, the diferent evolution paths of the involved sub-communities require a sophisticated concept for federated data space evolution. This keynote talk reports experiences of the ”Data Space Culture”, a lighthouse project of the German Chancellors Ofice aiming at investigating these issues and demonstrating and evaluating a suitable data ecosystem around four use cases in the fields of theaters, museums, music training, and networking local culture communities. We also discuss the potential synergies and interoperation challenges with the many other culture digitization initiatives in Europe and beyond.
ifelds, characterized by a wide diversity in digitalization. Some large individual players (e.g. the Deutsche Museum in Munich) have invested intensely in all kinds of digital upgrading of their artefacts, or (like the movie and music industries, or the Europeana cultural heritage sector) are organized by large-scale multimedia data collection and streaming platforms, or multi-year joint eforts such as the Europeana or the German Digital Library. IT infrastructure and limited access to their intended audiences. Moreover, the potential advantages through joint value creation through controlled B2B data sharing without giving up sovereignty of their own data rarely expoited. ture as one important application and innovation domain for the concept of data spaces, but other areas such as logistics, mobility, and manufacturing or even health have Bases (VLDBW’23) — Data Ecosystems (DEco), August 28 - September 0000-0001-6169-2942 (M. Jarke) been earlier in starting specific initiatives.
the Culture Department of the German Chancellor’s ofifce (BKM Bund) has therefore funded a large-scale experimental efort to set up a Cultural Data Space. The project is led by the German Academy of Science and
need for some important extensions of the emerging data space technologies, but also some important implications for governance and business models. In this keynote paper, we give a short summary of this initiative and its
The European Data Strategy therefore mentions cul- linkages to other data space application domains.
technical storage formats, bringing database technologies © 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License was semantic modeling and querying across the diferent
together with information retrieval as well as semantic networks and later linked open data.
In the 2010s, this original small-scale concept grew
into large muli-player systems in two dimensions. First, Intense political and technical discussions resulted in the
structured data warehouses of the late 1990s evolved to decision to strengthen and validate attractiveness and
Data Lakes which store raw data from many subfields or feasibility of a Culture Data Space by
sub-organizations, and allow data cleaning, detection of
related data, and sophistcated data analytics and machine
learning from many diferent perspectives, without big
upfront integration investment [
The second dimension emphasized decentralization of
data ownership and data usage control. Driven by the
importance of niche players and hidden industry
champions in Europe, an intense debate on data sovereignty
emerged, e.g. in the GAIA-X initiative [
A minimal data space system involves the following core functionalities [4? , 5]: • exploring the possibility of reusing structures and software components of a somewhat similar already operational data space, in the field of human mobility; not surprisingly, this turned out to be nontrivial, as this data space and even its the underlying base software was and is rapidly evolving. • identifying the adaption and extension needs of a broad range of cultural fields with specific highvisibility use cases, such that we could address the common core functionalities; • coping with the complex multi-organizational setting which on the one hand involves cultural institutions and creativity industries at multiple scales, and on the other hand regional interests ranging from local communities to state and federal to even European level.
To summarize, the data space technology enables the
formation and operation of communities of data
exchange and cooperative, secure value creation [
• standardized connectors as a kind of wrapper
gateway for the import and export of data to/from
the organizational information system of a
participant; thus, connectors can serve the technical
roles of data supplier, data requester, or both; the
units of data exchange are often digital shadows The four use cases were designed to analyse and
showof objects or activities which carry some value case specific opportunities and challenges for important
[
Museum exhibitions often face the challenge that only a few original works related to the theme – each often worth millions – are available locally, others must be borrowed and insured at enormous cost, or in digital form.
The reuse of artefacts and processes for later purposes
Organizational Implications
• a scaling of participant numbers over previously
studied data space applications by at least two
orders of magnitude;
• a culture-specific extension of access and usage
policies, including the embedding of existing
monetization organizations such as, in Germany,
VG Wort, VG Media, or GEMA for musical
performance rights;
• to avoid double work and unnecessary
inconsistencies, the creation of interoperability and
functional synergy with large public or private
data collections, such as the German and
European digital libraries„ public archives, and
media organizations, but also compatibilities with
the onging eforts towards domain-specific
FAIRcompatible research data infrastructures [
Space Support Center;
• methods and tools for semantic interoperability
among the many existing and forthcoming
metadata standards in the various cultural domains
but also to related data spaces such as Mobility,
Tourism, European Cultural Heritage, and the
like.
• in addition to the above-mentioned control-level
and metadata infrastructure, also culture-adapted
optimization for the actual data exchange and
value-added processing, including data-kind
specific storage, query,integration, and
personalization services in the highly heterogeneous setting,
following a logic-based approach as in [
These in part unique requirements imply careful think- data, and the value creation and fair value appropriation
ing about a suitable data space architecture. Among the [
The early data space applications, such as the German central approach, in particular aiming at federated broMobility Data Space, bundle all the core data space func- kers with intelligent metadata translation facilities. By tions – except the connectors of the participants in a the end of 2023, two of the four use cases will be in part single operational organization which provided the core operational, all of them fully a year later. In parallel, the services such as broker, contract, and identity manage- generic topics of advanced rights management, the govment. It seemed natural to start with such an already ernance structure of the federated data space ecosystem, operational infrastructure for the first experiments, and and the future operating companies, and the business define a growth path for the new requirements. models will be decided.
For data spaces with tech-savvy large industrial participants, such as the emerging catena-X data space for the automotive industry, this architecture was consid- Acknowledgments ered to limited, but also too intrusive since e.g. usage policies were intended to be enforced into the individ- This work was supported in part by the Beauftragte fuer ual connectors of participants, such that they felt to lose Kultur und Medien der Bundesregierung as a lighthouse some control over their own information systems. They project Datenraum Kultur within the digialization stratpreferred a solution in which they take full responsibil- egy of the German Federal Government. Special thanks ity for linking their internal system to connectors. The go to Georgios Toubekis who coordinated the complex connectors themselves have only a very basic function- use case analyses. ality, and the participants can add those of the formerly central services they need as so-called extensions to the References connectors. The open source development program for this efort is mostly performed in the EDC connector initiative of the Eclipse foundation by industrial and science partners including Fraunhofer.
However, this approach assumes high IT-/data management competence of at least the largest part of data space participants, which is totally unrealistic for most of the cultural sector. Moreover, diferent user groups and sub-dataspace communities progress at very diferent speeds, e.g. depending on whether they can invest their own money or need to rely on (initial) public support.
Even from the experiences within the current project, it seems unimaginable to think of a single governing body which could make suficiently quick and mutually acceptable decisions. We therefore opted for a third architecture which consists of several federated sub-dataspaces run by a culture domain, a very large data repository organization, a city or rural community, or perhaps by a creative industry organisation which would ofer similar services as the original IDS approach but e.g. federated brokers and contract support for data exchange across the boundaries of the subcommunity.
Our case studies have shown remarkable additional attractions and requirements for a Culture Data Space. We stress again that our understanding of such a data space focuses on the sovereign sharing of diferent kinds of