Development of performance monitoring system based on Alternate Marking methodology (RFC8321) and the Big Data Approach. Starting from a previous model, this work aims to improve some parts of the latter. In particular the thesis consists of two parts, one concerning the adaptation and the modification of a new probe to the preexisting model, and another part concerning the use of Apache Kafka to facilitate and improve the way in which collected data are sent. Tests about scalability and performance of the latter have been done using three servers having multiple probes and multiple streams. Conclusions on these tests were made in order to evaluate the validity of the proposed solution.