Executing Complex Calculations in the Cloud to Enable Real-Time Data Processing

Complex and time-consuming calculations usually cannot be processed on smartphones in a real time manner, as recent researches show. For these calculations a high-end mobile hardware setup is needed. This creates a high latency which is a drawback as the results cannot be visualised during the evaluation process upfront. This paper proposes a system for spreading the compute-intense calculations on high performance computing units to a cloud computing architecture by wrapping the complex calculations into native function calls and exposing them via a Representational State Transfer (REST) service. This system makes it possible to expose the Dynamically Linked Library (DLL) via Application Programming Interface (API) endpoints creating a reliable low latency Transport Communication Protocol (TCP) connection. Our paper shows that spreading the compute-intense calculations on high end computing units leads to reduction of computation time. This enables the user to reliably stream the results back in real time for further processing (e.g. visualisation).