Context Information Management across Smart City Knowledge Domains

This presentation summarizes ongoing work within the ETSI technical group ISG CIM to standardize and improve interoperability for discovery and exchange of context information between applications, IoT systems and "legacy" databases, particularly for Smart City use cases. The goal of the talk is to solicit comment on issues of scalability and interworking.

A Context Information Management system offers a clearing-house for publishing, discovering, monitoring and manipulating the data which has the right context for an application, in particular an application which did not itself generate the data. Data is usually interpreted within the "vertical" knowledge domain where it is collected, but taken away from its context, each piece of information is often nearly useless. Even worse, people and especially software programs/agents searching for useful information can only find useful data if the context is available i.e. published with the data in a usable format.

The need for interoperable context information management is growing fast due to (a) trends in eGovernment and Open Linked Data; (b) requirements on provenance information, which includes also licensing constraints; (c) privacy regulations requiring ability to discover/purge private information across multiple databases; (d) validation of reasoning chains using diverse data sources, particularly regarding liability for actions triggered by machine-reasoning systems

The presentation will provide a short status report on the work of ETSI ISG CIM (https://docbox.etsi.org/ISG/CIM/Open) and the problem space for context information exchange, followed by discussion time for key issues. The goal of the group is to identify an API which is fit-for-purpose to efficiently publish, discover and subscribe to information about (m)any kinds of things, for many kinds of applications which can range from recreational (a family travelling wants to detect "what are the important tourist sites near me") to safety-critical (a fire department needs to check risk of crowd panic at a major city event).

The key problem for the group is that the solution requirements are known to cross many research areas, with deep levels of complexity and accumulated expertise, and it is hoped SEMANTICS 2017 will provide contact with key experts in:
(a) practical assignment of unique identifiers which are robust against linking to multiple (non-permanent) information sources providing context information;
(b) efficient linking of (defined subsets of) data across distributed systems with disjoint administrations;
(c) efficiently enabling privacy control where required using fine-grained access control and/or encryption;
(d) controlling and monitoring usage by n applications to m data-sources for purposes of usage statistics or access optimization, including caching/paging/mirroring;
(e) validation and pruning of link chains;
(f) validation of data during the registration/publication process;
(g) real-world experience in cross-linking databases which were never designed for it!