This paper presents a description of a data acquisition IoT platform architecture which has been designed to provide two additional extensible mechanisms at server and device levels. From the device perspective we support implementations for two data acquisition modes: near real time and batch. From the server side we allow data analytics processes to be seamlessly configured and launched over the data acquired. Our approach pursues to reduce the efforts and cost of integration of new hardware devices. Along with the architecture description we provide an API definition, a platform reference implementation and its performance evaluation. Our main purpose is to guarantee integration, modularity, flexibility and extensibility to cover a wide range of applications and use cases.