Stream Documents Processing Invariance in Situation-Oriented Databases

Abstract

Data stream processing in situational-oriented databases (SODB) is discussed. SODB is a heterogeneous data integrator operating under the control of the built-in hierarchical situational model (HSM). The processed data is defined in the HSM as virtual documents (VD) mapped to real data stores. HSM specifies the loading of a VD data into Data Processing Objects (DPO), the data transformation in DPO, and unloading of processing results to a VD. Stream processing of large documents that do not fit entirely into RAM is discussed. A VD model involving multiple pieces of data that can be processed separately is considered. Portions of data are extracted from the VD input stream, processed in the DPO, and then sent to the VD output stream. Invariance should ensure the independence of the DPO model when the VD model changes. Invariance to different schemes streaming data is discussed. Algorithms for loading and unloading DPO are considered, ensuring invariance due to the hidden cyclical processing of data portions. The proposed solutions are illustrated with examples of processing XML and JSON documents.

Graphic Documents Parametric Personalization for Information Support of Educational Design Using Situation-Oriented Databases

“…In Fig. 2 shows the types of external data available for display: local and remote files, web services, archives, relational databases [16][17][18][19][20][21]. Data processing is performed using processing objects.…”