Accelerating smart eHealth services execution at the fog computing infrastructure

Abstract

Fog computing improves the execution of computationally intensive services for remote client nodes as part of the data processing is performed close to the location where the results will be delivered. As opposed to other services running on smart cities, a major challenge of eHealth services on the fog is that they typically span multiple computational activities performing big data processing over sensible information that must be protected. Using the capacities of current processors can improve the servicing of remote patient nodes. This paper presents the design and validation of a framework that improves the service time of selected activities at the fog servers; precisely, of those activities requested by remote patients. It exploits the capacities of current processors to parallelize specific activities that can be run on reserved cores, and it relies on the quality of service guarantees of data distribution platforms to improve communication and response times to mobile patients. The proposed approach is validated on a prototype implementation of simulated computationally-intensive eHealth interactions, decreasing the response time by 4x when core reservation is activated.