Information Technology Reference
In-Depth Information
middleware engineering and, in particular requires an approach that create and
maintain the models of the current networked systems and exploit them to reason
about the interaction of these networked systems and synthesise the appropriate
artefact, i.e., the emergent middleware, that enable them to interoperate. How-
ever, although the specification of system capabilities and behaviours have been
acknowledged as fundamental elements of system composition in open networks
(especially in the context of the Web [8,16]), it is rather the exception than the
norm to have such rich system descriptions available on the network.
This paper focuses on the pivotal role of learning technologies in supporting
Emergent Middleware, including in building the necessary semantic run-time
models to support the synthesis process and also in dealing with dynamism by
constantly re-evaluating the current environment and context. While learning
technologies have been deployed effectively in a range of domains, including in
Robotics [26], Natural Language Processing [20], Software Categorisation [25],
Model-checking [22], Testing [12], and Interface Synthesis [2], and Web service
matchmaking [15], this is the first attempt to apply learning technologies in
middleware addressing the core problem of interoperability.
This work is part of a greater effort within the
Connect
project
1
on the
synthesis of Emergent Middleware for GMES-based systems that are represen-
tative of Systems of Systems. GMES
2
(Global Monitoring for Environment and
Security) is the European Programme for the establishment of a European ca-
pacity for Earth Observation started in 1998. The services provided by GMES
address six main thematic areas: land monitoring, marine environment moni-
toring, atmosphere monitoring, emergency management, security and climate
change. The emergency management service directs efforts towards a wide range
of emergency situations; in particular, it covers different catastrophic circum-
stances: Floods, Forest fires, Landslides, Earthquakes and volcanic eruptions,
Humanitarian crises.
For our experiments, we concentrate on joint forest-fire operation that involves
different European organisations due to, e.g., the cross-boarder location or crit-
icality of the fire. The target GMES system involves highly heterogeneous NSs,
which are connected on the fly as mobile NSs join the scene. Emergent Middle-
ware then need to be synthesised to support such connections when they occur.
In the following, we more specifically concentrate on the connection with the
Weather Station NS, which may have various concrete instances, ranging from
mobile stations to Internet-connected weather service. In addition, Weather Sta-
tion NSs may be accessed from heterogeneous NSs, including mobile handheld
devices of the various people on site and Command and Control —C2— centres
(see Figure 1). We show how the learning techniques can serve complement-
ing the base interface description of the NS with appropriate functional and
behavioural semantics. It is in particular shown that the process may be fully
automated, which is a key requirement of the Emergent Middleware concept.
Search WWH ::
Custom Search