The new Internet is evolving into the vision of the Internet of Things, where physical world entities are integrated into virtual world things. Things are expected to become active participants in business, information and social processes. Then, the Internet of Things could benefit from the Web Service architecture like today's Web does; so Future service-oriented Internet things will offer their functionality via service-enabled interfaces. As demonstrated in previous work, there is a need of considering the behaviour of things to develop applications in a more rigorous way. We proposed a lightweight model for representing such behaviour based on the service-oriented paradigm and extending the standard DPWS profile to specify the (partial) order with which things can receive messages. To check whether a mashup of things respects the behaviour, specified at design-time, of composed things, we proposed a static verification. However, at run-time a thing may change its behaviour or receive requests from instances of different mashups. Then, it is required to check and detect dynamically possible invalid invocations provoked by the behaviour's changes. In this work, we extend our static verification with an approach based on mediation techniques and complex event processing to detect and inhibit invalid invocations, checking that things only receive requests compatible with their behaviour. The solution automatically generates the required elements to perform run-time validation of invocations, and it may be extended to validate other issues. Here, we have also dealt with quality of service and temporal restrictions.

run-time verification, composition, mashup, behaviour, service-oriented things, internet of things, web of things, complex event processing, mediation patterns