Implementation and Evaluation of a Prototypical Service for the Processing and Provision of Clinical Trend Data in a Distributed System

The ISO/IEEE 11073-SDC standard defines architecture and protocol on interoperability between manufacturer-independent medical devices. In this paper, a prototype implemented this standard to process clinical trend data using Apache Kafka for data storage and GraphQL for its interface. The proposed prototype demonstrated real-time and historical data connected to 100 simulated devices. In the evaluation, the processing time of trend during high-load setting has occasional outliers, however its median is relatively low. A single historical trend query with varied payload and real-time subscriptions were effectively executed, meanwhile simultaneous queries from multiple devices caused high latency due to resource contention.