Intrinsically, the Internet of Things will be « non deterministic », complex and various.
Actually, the emerging technologies and concepts (sometimes “standards”) that currently makes it up will accordingly lead soon to both :

So far, the ongoing standardization models (eg : EPCGlobal) are still apprehending deterministic value chains or processes even though some probabilistic approaches (eg : BRIDGE project - are intending to extend the scope of available means to address complex issues such as lookup services (Discovery Services, IETF-ESDS, etc).

Actually, both “B-to-B”, “B-to-C” or “structured EDI” schemes are still the underlying framework in many cases.

However, through inherent nature, the Internet of things will be made of heterogeneity, diversity, exceptions and subsidiarity.

Therefore, the existing syntaxic methods (eg : structured EDI) will not be sufficient to give anyone sufficient capability to communicate, exchange or interoperate with any other one over this “Internet” (“anyone” can be objects, web services, HMI, etc.).

In such an environment, the use of Semantic approaches will, possibly, give a “meaning” to every kind of message and will accordingly facilitate communications whoever will be the actors, even though they have no prior interchange agreement.

Next generation of EDI systems could therefore be based on technologies similar to “Text mining” instead of existing schemes (ODETTE, EANCOM, etc).

Last, but not least, it will help as well to avoid any mono-model approaches that creates monopolies such as the existing ones….

Philippe GAUTIER, june 2008