Avaliação dos brokers Kafka e Apache Flume no contexto de big data

Abstract

The data volume is getting bigger and different solutions have been proposed to deal with a large volume of data. This paper presents a performance evaluation of the Kafka and Flume brokers in the context of Big Data. We proposed a benchmark to evaluate the brokers with simulated data and a real application. When we use previous stored data the broker Flume performed better than Kafka and when the streaming is used to generate data Kafka has a better performance.