SeCEE: Edge Environment Data Sharing and Processing Framework with Service Composition

A centralized computing paradigm, such as cloud computing, cannot satisfy the explosive growth in the amount of data and computing needs. Therefore, a computing paradigm for the edge environment is proposed to enable real-time data analysis with large volumes of data by decentralizing heavy computational loads and reducing the consumption of network bandwidth. Data ownership can provide substantial commercial interests for data owners. However, traditional data processing exposes data on the Internet and incurs the risks of data value reduction and privacy issues. By applying the computing paradigm in the edge environment to data sharing and processing, people can build data processing applications without providing the whole original dataset. However, existing work on lightweight methods of building applications and decomposing computation tasks is still lacking. In this paper, we present SeCEE, a framework for data sharing and processing in the edge environment. This framework utilizes geographically distributed datasets to analyze data without programming and comprises (i) a hierarchical task network-based approach that describes datasets and corresponding services from different stakeholders, on the basis of which the features and relationships among datasets and services are recorded; (ii) a service composition method that instantiates an abstract process model for multiple data flows in a dynamic environment; and (iii) an execution engine that coordinates the computing process by dispatching computing tasks to edge servers and collects results for combination and further processing. A case study of a data processing application for electronic toll collection demonstrates the effectiveness of the proposed framework.