Regresar

Service Response Time Estimation in Crowdsourced Processing Chain

Abstract:

Self-organized platforms to support Citizen Observatories need to build data processing chains based on the study’s goal where the non-functional (e.g., response time) requirements are as important as the functional requirements. In this research, a method to estimate the average response time of data services commonly found in crowdsourced data processing chains by using basic statistics from the data distribution of previous campaign is proposed. The method is evaluated using 18.512 registers of map-matched trip segments collected in a citizen mobility campaign gathered by 310 devices. Results show that the proposed method report an error rate between 5.55% and 12.55% when the transport mode is not considered, and between 5.12% and 9.41% when the transport mode is used.