Page 122 - ITU Journal - ICT Discoveries - Volume 1, No. 2, December 2018 - Second special issue on Data for Good

ITU JOURNAL: ICT Discoveries, Vol. 1(2), December 2018
• infrastructure and transactions auditing for
performance and disputes.
5.3.3 Matching of parties (brokerage)
Brokerage and third party services are typical
components of traditional market exchange
services, however it is not trivial to enable such
services over data exchanges. The following are
essential brokerage services to enable effective and
trusted data exchange:
• directory and registry services to enable
findability of data and services, including
data providers, target data applications and
typical data uses, conditions, etc.;
• data filtering and quality assessment
services, data processing API, value added
service/ product offerings;
• automated/assisted contracts negotiation
and data “value/property”
transfer/exchange, customer support.
5.3.4 Data catalogues to keep everything
together
A data catalogue (or Directory) is an essential
component of the intended data market and data
exchange infrastructure that can be distributed and
operated in centralized hierarchical and federated
modes. Catalogue service should enable both
hosting metadata information and
historical/provenance information about data sets
and data transformation. The following summarizes
the data catalogue properties and services:
• cataloguing data sets
• cataloguing data operations
• metadata catalogue
• searching
• data curation
• data quality assessment and data
categorization
• linking data properties and applications
• recommendations and relationships
• data sets evaluation
• data access policies and API
• usage metadata
• lineage/provenance
• integration and interoperability
100 © International Telecommunication Union, 2018
• supporting data pricing models (current and
future)
• data visualization (properties, usage, quality)
• security
• compliance.
5.3.5 Data exchange protocol
Data exchange protocols are important components
of future data markets and data exchange
infrastructure. Data exchange protocols can be
provided as a part of the data exchange (as a data
market functional component) and use a layered
model benefiting from reliable and secure modern
Internet protocols, while deploying specific data
exchange protocols as upper layer protocols.
Modern network and cloud virtualization
technologies allow the building of program able
virtualized networks that can be provided as part of
virtual private clouds (VPC) and support data
applications that produce and consume data and
involve data exchange internally inside secure VPC
and with external parties.
Recent developments in data descriptions and data
management introduced a number of mechanisms
that can be used for effective and consistent data
exchange protocols and applications: persistent
data identifiers (PID), data factories, metadata and
data types registry, data annotation and data
discovery mechanisms, all these are part of the
Research Data Alliance (RDA) outputs and
recommendations [33].
6. CONCLUSION
The presented paper is intended to provide a basis
for further wider discussion on defining data
properties as economic goods and building future
open data market and data exchange infrastructure
by involving concerned communities, in particular
those that the authors are actively involved with.
These are Research Data Alliance, Industrial Data
Space, European Open Science Cloud, and National
Institute of Standards and Technologies (NIST) Big
Data Working Group. The proposed definition of the
STREAM data properties is based on extensive
study of related works and authors’ experience in
building modern big data infrastructure and cloud
based services.