Keywords: Access, Data Service, Service Design

Affiliation: University of Maryland, Department of Computer Science

Area of Application

The BD-DS modelling method aims at providing an approach (procedure, structure and related mechanisms/algorithms) to "Design Semantics on Accessibility in
Unstructured Data Environments". The modelling method establishes conceptually the idea of "Data Services" and is supported by a modelling environment for the design of the concept.. The idea and objective of the project is shown in the below figure.

The modelling method is positioned as a mediation layer between functionality on one hand (any kind of business or application functionality) representing data requirements, and the data providing layer on the other hand. The data service is functionally responsible to align the demands with provisioning, allowing the design of access techniques, query and composition mechanisms and result representation.


In any kind of data-driven implementation, the need to access data in a transparent and flexible way has become increasingly important over the last years. This is a consequence of the technological advances on infrastructure level, allowing us to store and process larger volumes of data, with a higher more complexity/variety and modification speed on structure as well as content level close to real-time. Any device, tool or interaction we perform currently produces data that is stored in a structured or un-structured way and could potentially be used to create advanced insight.

The work performed leading to the tool BD-DS aims to establish the concept of “Data Service” as a model-based approach to match the data needs of services on one hand with existing, (un-)structured data environment. The approach is conceptualized as a domain-independent modelling method enabling developers to semantically design access mechanism and algorithms, combine available data services according to functionality requirements and enrich retrieved data with meta-data and provenance information.

The concept is realized as the modelling method “Big Data – Data Service (BD-DS)” to semantically map data sources of different kinds, formats and representations with data demands on application level. The modelling tool allows the user to a) model the access and Extract-Transform-Load (ETL) of unstructured (and structured) data, b) flexibly adapt to schema changes using conceptual models, and c) result in the formation of novel aggregate data services.