Apache Kudu has been included in FEEP IoT & Big Data Platform Sofia2 release 3.4 as Real Time Database (RTDB). Kudu is an open-source solution that supports mixed workloads: real-time and analytics through an efficient scanning mechanism over a single storage layer.
Using Kudu as RTDB, the information provided in the ontologies will be mapped to tables, although inserting and obtaining data with JSON format is also available.
This implementation has been developed using the Hive JDBC access, which can stablish connection with Impala and Impala-Kudu interconnection.
With Kudu, the difference between RTDB and HDB has been disolved, allowing storage on a single layer of storage and simplifying the architecture of mixed applications.
For installing Sofia2 with Apache Kudu, you only have to modify some configuration parameters to enable the DAO.
With these features, the creation of the ontologies will be available with “Step by Step Creation”:
Once an ontology is created, it will be not possible modify, create or delete the fields contained in it.
The sending and reception of data is independent of the chosen data model. In case of Kudu, there is a manager that is responsible for transforming the incoming messages to a proper format. Here we can see an example of this:
These data have been saved in Kudu as:
From de RTDB & HDB console in Sofia2 we can query on any ontology and choose the data view mode: JSON format or table