RELYonIT will enable the cost-effective construction of IoT applications that provide performance guarantees in hostile environments. To this end, RELYonIT devises environmental models that capture how environmental properties vary over time and platform models that capture how these environmental properties affect the operation of a hardware platform. These models typically have parameters and we will develop techniques and tools to learn these parameters for a given platform and for a given deployment environment.

Existing protocols will be optimized and where necessary new protocols are designed that provide performance guarantees for a given deployment environment and hardware platform by exploiting knowledge from the respective environmental and platform models.

A specification language will be defined to allow a user to specify dependability requirements for a given application. Given such a specification of dependability requirements, techniques and tools will be provided to automatically select protocols and their parameters such that the performance requirements are met for a given deployment environment and hardware platform or the infeasibility of these requirements will be detected. If runtime assurance detects that environmental properties have significantly changed, the parameters of the protocols will be automatically adapted to reflect the new environmental model.

A fundamental requirement of RELYonIT is the availability of an experimental infrastructure with a large number of different hardware platforms that can be exposed to a wide variety of environmental conditions. While FIRE experimental facilities such as WISEBED or SmartSantander meet these requirements, it must be possible to repeat an experiment under identical environmental conditions. Existing FIRE testbeds will therefore be extended with the ability to record and playback the effects of environmental conditions.

Demonstration and Use

The work in RELYonIT is driven by real-world scenarios that are of interest to our end-users ACCIONA and WorldSensing. One of the candidates for a final demonstrator is a smart parking solution that allows users to find the nearest free parking spot using a smart phone application. Sensors embedded into the pavement detect the occupancy of parking spots. These sensors are exposed to extremely variable environmental conditions, yet the system must meet latency and lifetime requirements.

Scientific, Economic and Societal Impact

RELYonIT has a dual contribution and impact. Firstly, existing FIRE experimental facilities are enhanced with the ability to record and playback realistic environmental effects. Compared to previous solutions that used custom hardware, our software-based approach is more cost-efficient and easily scales to a large number of FIRE experimental facilities with many nodes each. The environment-dependent performance and thereby the feasibility of IoT applications can be predicted prior to deployment by running it on an enhanced FIRE experimental facility.

RELYonIT, secondly, uses this enhanced infrastructure to provide a dependability framework for the IoT. With that, a large class of dependable IoT applications that could not be realized before becomes technically feasible. The time and cost needed for developing dependable IoT applications will be substantially reduced, thereby making certain dependable IoT applications economically feasible. Safety and comfort of a future IoT-enabled world will be substantially improved.