The Design Principles and Practices of Interoperable Smart Spaces - Advanced Design Approaches to Emerging Software Systems - page 36

cable for client-server solutions, commonly used
in indoor spaces. For example, Smart-M3-OSGi
integration was applied to the smart maintenance
of public buildings demonstrator (Manzaroli
2010). NoTA is a novel service architecture for
networked embedded systems applied to a personal
space demonstrator (Luukkala 2010). IOP based
on Smart-M3 and Web Services are applied to a
smart city pilot, still under development.

well suited to modeling real-world phenomena,
including entities, events, interactions and seman-
tic connections in smart environments. Whenever
greater expressivity is needed, simple RDF can be
supplemented with OWL 4 constructs, constantly
keeping attention at avoiding an unnecessary
complexity in ontology definition. OWL provides
further modeling primitives to describe properties
and classes: among others, the relations between
the classes, cardinality, equality, richer typing of
properties and the characteristics of properties
(e.g. symmetry), and enumerated classes.
Given the wide spectrum of targeted smart
space applications, a crucial issue in ontology
development is how to develop a set of suitable
ontologies based on a common data model that
can be specialized to each application field. This
problem was not only a modeling issue, but also
had strong performance implications since the
loading of an oversized set of ontologies into a
SIB or having it processed by a KP would af-
fect both the efficiency and scalability of SSA.
Therefore, we adopted a three-layered approach
for the ontology development:

ONTOLOGY DEVELOPMENT

Smart space applications are based on three
“abilities” that make them 'smart': i) an ability to
understand the situation where the application is
used and by whom, ii) an ability to interpret the
semantics of shared information, and iii) an ability
to tolerate uncertainty at development time and
at run time. In particular, the first and the third
point embody the concept of context-awareness,
while the second relates to information interop-
erability. In this section, we will focus on the
semantic modeling of data to achieve information
interoperability. Thereafter, we will discuss how to
develop context-aware smart space applications.
Several interoperability issues are raised by
the heterogeneity of devices and software which
are already in place in physical environments. To
reduce development and deployment costs and to
maximize the reuse of existing applications, our
methodology tackles those interoperability chal-
lenges at the information level via the adoption
of proper semantic technologies. The primary
idea is to provide smart applications with highly
interoperable and shared information spaces that
maintain sensed data and information on currently
available resources and services. The content of
shared spaces is openly understandable and largely
re-usable thanks to the exploitation of lightweight
semantic technologies, first of all, Resource De-
scription Framework 3 (RDF)-based ontologies to
describe simple relationships between represented
entities. The data model of RDF is generic and

• a foundational ontology layer to provide

the base concepts and relations needed for
a real-world description model (e.g., the
concepts of an event or person);

• a core ontology to provide the concepts

and relations common to all SSAs (e.g.,
the concepts of a smart device or quality of
information);

• a set of domain ontologies , specific to

each application domain, to provide the
concepts and relations describing targeted
scenarios (e.g., the concepts of tempera-
ture and building in a home maintenance
scenario).

Each layer is built on top of the previous one;
in that way, the higher level exploits and extends
the lower one. The DOLCE ontology (DOLCE
2010) was selected for the foundational ontology