Abstract: This article presents a data management solution based on the data distribution service (DDS) communication model. The basic DDS communication model consists of a unidirectional data exchange where applications that publish data ‘push’ the relevant data, which is updated to the local caches of co-located subscribers to the data [1]. DDS has no specified center node to forward data packets and maintain the communication data. This type of publish-subscribe (P/S) model presents integrity and consistency challenges in data management. Unlike peer-to-peer (P2P) distributed storage, DDS applications have a hard real-time environment and fewer data features, and the core problem is ensuring the integrity and consistency of data in distributed systems under this hard real-time environment. This article begins with a brief introduction of the communication model used by DDS, then analyzes persistent data management problems caused by such model, and provides an appropriate solution to these problems. This solution has been implemented in a prototype system of the real-time service bus (RTSB) of Tsinghua University.

Key words: DDS, pub-sub, consistency, load balancing