Service entities model for the internet of things: A bio-inspired collaborative approach

This paper provides a model for internet of things (IoT) establishment. Today and for the coming decades we see rising interest in interconnecting `Things' as we did in the past two decades when we interconnected PCs and servers to build the called Internet. Smart things interconnection is more general and complex due to different challenges: heterogeneity of devices, diversity of protocols, variety or none established standards, self-manageability, self-organization, dynamic architectures, mobility, intermittent availability, distributed computing, security and privacy. Awareness of these challenges is increasing but dedicated solutions are still at the starting-point. Current proposal is a model allowing smart things interconnection and interoperability. It is mainly based on `service' concept but can bind peer models like `object' concept and standard-based architectures. It is, also, a tentative to learn from biological entities and real environment. `Things' in the IoT are at the origin smart electronic devices with some communication capability. These devices, in our concept, are not considered as unique entity but fragmented into a set of services independently of their architecture, functionality and content. Services in IoT are much more than simple data/document exchange, they are expanded, but not limited, to real world actions, metrics and events.

I. Introduction

Smart things interconnection is rapidly ramping-up [12]. Many or complex entities are reusing existing internet infrastructures like internet protocol (IP), Web servers, short range area network protocols and established architectures. But due to the complexity and cost involved by such solutions, it is not possible to reuse it for any (thing) interconnection to build the formal Internet of things (IoT). There are many parameters that make the reuse of existing standards insufficient, too expensive or simply inadequate. Some of these parameters are covered by key aspects of: interoperability of heterogeneous systems, mobility, ubiquity of processing, interconnectivity through heterogeneous protocols, standards and networks, adaptability, interactivity and autonomy. This is a real paradigm that requires long-term research and efforts from both academicals and industrials [4]. Added to this, involved technology spectrum is much larger than usual internet needs. It is difficult to subdivide such problem into OSI-like layers and focus on each layer; it requires the knowledge of all involved entities: hardware, networking, data formatting, data packaging, software, middleware and end-users experience. Use of rigid standard cannot resolve the problem; we need to think about other ways and techniques! If standards setup a detailed framework and enforce others to adhere; Internet of things can be considered slightly on the opposite side: use many standards but insure the interconnectivity and interoperability. This expresses a need to have two main features in the future smart things: autonomy regarding internal functionality where standard can be applied and high abstraction level toward external environment for interconnectivity and interoperability. Due to the diversity and dynamism of contexts, application and environment; future devices should integrate other dimensions that consist of learning and thinking. It is the era of embedded intelligence as a natural evolution of embedded systems [4].