IoT applications consist of diverse elements (​aka Things) including resource-constrained/rich devices with a considerable portion being mobile. Such devices demand lightweight, loosely coupled interactions in terms of time, space, and synchronization. IoT middleware protocols support one or more interaction types (e.g., asynchronous messaging, streaming) ensuring Thing communication. Additionally, they introduce different Quality of Service (QoS) features for this communication depending on the available computing and networking resources. Things employing the same middleware protocol interact homogeneously, since they assume the same functional and QoS properties. However, various IoT middleware protocols have emerged over the last few years, resulting in isolated islands of Things. It is then hardly possible to leverage those islands together within a single system due to the heterogeneity of the IoT protocols across layers, from the networking up to the applications. The heterogeneity of things further results in highly diverse quality of service, from the quality of​ ​the​ ​sensed​ ​data​ ​to​ ​that​ ​of​ ​its​ ​delivery.

Our aim is to enable the design and development of emergent mobile systems, which are dynamically composed from available Things that are networked in the given environment. Our approach relies on software architecture abstractions, including novel connectors for multi-layer in-network processing, combined with supporting formal models for composing and adapting the system's communication architecture at runtime. A key goal is so to ensure the resilience of the enacted sensing and actuation systems. This involves improving the reliability of the data submitted by the low-cost and miniaturized Things in order to deliver a global service that achieves the desired data quality. We address this challenge through in-network collaborative processing for the multi-party correction of observations. Further, the communication protocols implemented by our connectors must ensure that Things may coordinate and communicate so that the overall IoT-based systems delivers a sufficient quality of service to the end-user. To ensure so, we build upon the runtime analysis of the connectors’​ ​​ ​end-to-end​ ​QoS. For the sake of validation, we design our approach by considering more specifically emergency​ ​response​ ​scenarios.

In the above context, the team’s overall objective is to provide a middleware solution enabling adaptive communication protocols for the effective composition of heterogeneous networked Things, which will be assessed in the context of emergency-response​​ ​​scenarios. Toward​ ​this​ ​goal,​ ​our​ ​research​ ​subdivides​ ​into​ ​the​ ​following​ ​complementary​ ​themes:

T1​​ ​​-​​ ​​Adaptive​​ ​​communication​​ ​​architecture​​ ​​for​​ ​​resilient​​ ​​sensing​​ ​​and​​ ​​actuation.
To support high scalability and interoperability in IoT systems, we study a communication architecture based on the flexible ​publish/subscribe (pub/sub) interaction paradigm. To deal with the large volume of data and peers that span a wide-area, we will elaborate a set of independent, communicating pub/sub brokers, forming a broker overlay. By relying on this architecture, peers can access the system through any broker using an existing IoT middleware protocol. In this setting, IoT devices can be considered as publishers; and actuation devices, software programs, and end-users as potential subscribers. Peers can be highly heterogeneous in terms of protocols, data formats and device platforms. We thus design protocol translation techniques for converting between data and primitives used by different IoT middleware protocols, by relying on our previous experience. Finally, we aim to utilize new networking technologies to design a​ ​network-aware​ ​and​ ​adaptive​ ​pub/sub​ ​architecture.

T2​​ ​​-​​ ​​Stochastic​​ ​​models​​ ​​for​​ ​​analysis​​ ​​and​​ ​​runtime​​ ​​decision.
In this research theme, we investigate a dynamic adaptation mechanism (or self-adaptation) in our pub/sub IoT deployments by relying on stochastic models. In particular, we aim to design and develop a logically-centralized Broker Controller Service (BCS) that gathers state information about the components within the broker networks (e.g., peers connections, data flow rates) to intelligently manage device and resource provisioning according to administrative policies. The BCS will perform tasks, such as provisioning brokers, monitoring, controlling the flow of data and QoS-aware adaptation. The aim is to dynamically manage the QoS and resource characteristics and needs of the heterogeneous peers and their brokers. Specifically, to meet QoS policies we rely on queueing theory for the modelling of end-to-end delays inside the broker network. Based on the estimated delays, adaptation will require​ ​taking​ ​ appropriate​ ​actions​ ​from​ ​BCS.

T3​​ ​​-​​ ​​Cross-layer​​ ​​in-network​​ ​​collaboration​​ ​​protocols​​ ​​to​​ ​​enhance​​ ​​the​​ ​​overall​​ ​​QoS.
This research theme is concerned with the study of coordination protocols across Things so that they collaborate toward the goal of the system they contribute to, while both enhancing the quality of the delivered data and optimising the consumption of the network resource. Bearing in mind the emergency-response scenario target, we will concentrate on protocols that contribute to: (i) enhancing the connectivity of the Things, and (ii) enhancing the quality of the observations delivered by the things, through the collaboration of Things and in a way that​ ​enhances​ ​the​ ​overall​ ​system​ ​reliability.

T4 - Supporting middleware solution that embeds the models and implements the adaptive​​ ​​protocols.
Our contributions to the above Themes T2 and T3 will lead to the implementation of related prototypes, which will be integrated within a middleware solution adhering to the communication​ ​architecture​ ​designed​ ​as​ ​part​ ​of​ ​T1.

T5​​ ​​-​​ ​​Evaluation​​ ​​using​​ ​​focus​​ ​​emergency​​ ​​scenario​​ ​​use​​ ​​cases.
We will evaluate the proposed solution in the context of emergency response scenario use cases, which we will elaborate in more detail in the course of the second year. The work will leverage the experience of the UCI team and their connections to first responder organizations​ ​at​ ​community​ ​and​ ​national​ ​levels.