The large volume of data coming from a variety of sources and in various formats, with different storage, transformation, delivery or archiving requirements, complicates the task of context data management. At the same time, fast responses are needed for real-time applications. Despite the potential improvements of the Smart City infrastructure, the number of concurrent applications that need quick data access will remain very high. With the emergence of the recent cloud infrastructures, achieving highly scalable data management in such contexts is a critical challenge, as the overall application performance is highly dependent on the properties of the data management service.
We argue that potentially large benefits can be obtained by using versioning to improve applications’ data access performance under heavy concurrency.
This is achieved by exploiting the inherent parallelism of data workflows efficiently, through a decentralized management, asynchrony, fault tolerance, and by allowing users to explicitly control written data layout such that it is optimally distributed for reading/writing. Real-time processing address the aspect of data locality, data processing depending of where data are stored and the overhead added by data transfers will add penalties in applications’ performance.