Internet of things scalability: Analyzing the bottlenecks and proposing alternatives

Internet of Things, or IoT, is a new revolution of the Internet, supporting a connection between objects and humans yet more strong. This area is known by the huge quantity of data that can be generated from the RFID sensors at realtime. On managing this data, the IoT middleware introduces some time constrains regarding both accessing applications and filtering and storing data actions. Thus, middleware scalability is a key requirement for supporting the current growth of IoT. In this context, this article firstly presents μIB - Micro Internet of Things Benchmark - that aims at evaluating the scalability of EPC-compliant middlewares. μIB combines data from tags, RFID readers, sequential requests and number of threads to compose an evaluation setup to assess EPCIS and ALE components. We used this benchmark to evaluate Fosstrak, which is one of the most used IoT middlewares, in a single server. The tests emphasized that the exhaustion in one component influences the performance of the others directly. Following this research line, this article also describes the first ideas on redesigning the IoT middleware using techniques such as cloud virtualization and elasticity, service decomposition and multithreading programming.