A3ME – Device-Agent based Middleware for Mixed Mode Environments

• Main Author: Herzog, Arthur
• Format: Others
• Language: German; en
• Published: 2016
• Online Access: https://tuprints.ulb.tu-darmstadt.de/5235/7/2016-02-03_diss_arthur_herzog.pdf; Herzog, Arthur <http://tuprints.ulb.tu-darmstadt.de/view/person/Herzog=3AArthur=3A=3A.html> (2016): A3ME – Device-Agent based Middleware for Mixed Mode Environments.Darmstadt, Technische Universität, [Online-Edition: http://tuprints.ulb.tu-darmstadt.de/5235/ <http://tuprints.ulb.tu-darmstadt.de/5235/> <official_url>],[Ph.D. Thesis]

The Internet of Things describes a vision and a process, that partially already takes place, in which the things become interactive by being provided with minimal computing power and communication capability to support the people in their tasks. These smart things or devices are nowadays prevailingly organized in island solutions and are often not able to interact across the border of their own island. As the amount of electronic devices we are surrounded by increases every day not only in numbers but also by their variety, enabling interoperability among these devices became crucial. Mixed Mode Environments (MME) refer to networks composed of very different kinds of devices, which are distributed among various physical environments and communicate with each other using different communication technologies. The single nodes in the network can be sensors, actuators, robots, unmanned vehicles (UV), computers, user interfaces, smartphones, etc. All those devices have their specific capabilities and constraints. Many of these devices are manufactured by different companies and use different software and operating systems. Those devices can communicate with each other by wire, radio, light, sound or other transmission medium. For each of these transmission media many different communication technologies exist, which use different protocols, frequencies, encodings, etc. Nowadays application developers have to deal with the before mentioned heterogeneity when developing a new application for such a network. Each time when a new kind of node appears the application has to be adjusted to deal with the new hardware. Middleware is a way to avoid this direct interaction of applications with the different hardware and software of the devices. So middleware has to abstract over all the different devices, their capabilities and communication technologies and to offer the applications uniform interfaces to interact with those. The Device-Agent based Middleware for Mixed Mode Environments (A3ME) framework developed in this dissertation enables interoperability among different nodes without the need of adjustments each time new hardware is introduced. Our approach offers an abstraction for the different hardware: it sees all the different nodes in the network as independent entities, we call them device-agents. These device-agents know the capabilities and constraints of the respectively represented device and represent those in a neutral format developed in this dissertation. This neutral representation is independent of the technologies used on the represented device. Depending on the capabilities of the represented device a device-agent offers services to other agents and can also use services of other device-agents. The complexity of device-agent running for example on a small sensor node and on an UV can vary considerably. Thus a sensor-agent might be only capable to measure the current temperature and to send it to someone who is interested in this data, whereas the UV-agent can move through the environment, collect the data from the sensor-agents, aggregate, evaluate the collected data and use the gathered information further in its decision making process. Communication between different nodes is defined technology independently and can be applied to different communication technologies. Descriptions of the devices and their capabilities are based on a common classification developed in this work. This allows interactions based on capability-classes. All interactions are message based and each message belongs to a specific performative, which corresponds to the type of action the message represents. The structure of the messages is defined using Abstract Syntax Notation One (ASN.1) allowing to define their content dynamically. The use of ASN.1 Packed Encoding Rules allows to encode the messages in a very byte-length efficient way. This framework offers the basis to enable interoperability between heterogeneous devices directly and among various island solutions for electronic devices, which exist in different areas like multimedia, personal communication, smart home, smart office, connected car, etc.