Safety-focused deployment optimization in open integrated architectures

Open Integrated Architectures like AUTOSAR or IMA enable a flexible deployment, which can potentially help to reduce the number of computer platforms in a distributed embedded system, and therefore reduce weight, energy consumption and costs. Finding a beneficial deployment is, however, a complicated, multi-criteria optimization problem. One criterion that requires exceptionally careful examination is safety, since an adverse deployment can compromise system safety and inflict significant costs. In this paper we present a technique that assists the developer in optimizing a deployment from the safety perspective. The technique consists of two metrics to evaluate the safety-related costs of a deployment that have been tested and evaluated in an industrial context using a genetic algorithm. System developers can use these metrics to evaluate and optimize a specific deployment with respect to safety.