Recently, Internet of Things (IoT) attract attention. The authors are developing a cloud platform for IoT applications. The IoT cloud needs to deal with various types of data and large data sets depending on applications and purpose of use. That is, the IoT cloud necessarily includes heterogeneous data stores in a mixed manner. For example, relational databases and NoSQL databases have different connection methods and query languages. This configuration complicates the system design and increases the development cost. This paper presents a configuration method of data access component (DAC) that absorbs the connection method and the query language differences among data stores. This allows us to develop IoT applications without worrying about data store differences and later replacements. In the implementation, we used specific DACs optimized for specific data stores and a multi-purpose DAC Apache MetaModel. With a large scale data set of more than one million records under most configurations, the response time for various kinds of queries are less than 1 second.
[1]
Ramon Lawrence,et al.
Dynamic Database Integration in a JDBC Driver
,
2005,
ICEIS.
[2]
José A. Blakeley,et al.
Anatomy of the ADO.NET entity framework
,
2007,
SIGMOD '07.
[3]
Akihito Nakamura,et al.
Evaluation on Information Model about Sensors Featured by Relationships to Measured Structural Objects
,
2016,
IoT 2016.
[4]
Elizabeth J. O'Neil,et al.
Object/relational mapping 2008: hibernate and the entity data model (edm)
,
2008,
SIGMOD Conference.
[5]
Luca Cabibbo,et al.
Managing Inheritance Hierarchies in Object/Relational Mapping Tools
,
2005,
CAiSE.