A QoS-aware distributed Message Broker for the IoT.
The Zefxis nessage broker is designed to support the deployment of large-scale, heterogeneous and dynamic IoT applications. These consist of diverse Things including both resource-constrained/rich devices with a considerable portion being mobile. Zefxis adopts a pub/sub interaction paradigm that enables decoupling IoT devices in both time and space and supports distributed applications spanning a wide-area through a set of independent, communicating brokers. Zefxis supports efficient delivery of data from IoT sources to relevant subscribers while considering several QoS semantics and requirements related to the mobile IoT.
Queueing Network (QNs) offer a simple modeling environment, which can be used to represent various application scenarios, and provide accurate analytical solutions for performance metrics (e.g., system response time). We use QNs to model the performance of middleware protocols found in IoT, which are classified within the pub/sub interaction paradigm – both reliable and unreliable underlying network layers are considered. We also consider QoS semantics for data validity, buffer capacities as well as the intermittent availability of mobile IoT devices. We then introduce a pub/sub mechanism for probabilistic event dropping by considering the devices’ intermittent connectivity and QoS requirements. The consequent PerfMP performance modeling pattern may be tailored for a variety of deployments, in order to control fine-grained QoS semantics.
We showcase the application of our analysis in concrete scenarios relating to Traffic Information Management systems, that integrate both reliable and unreliable participants. We currenlty develop a prototype implementation of the zefxis message broker.
Useful links and references
- The Zefxis performance modeling pattern has been used in the ACHOR international research project to allocate broker nodes on Cloud.