This platform reduces execution time by 92.77% in SQL queries across multiple data sources, suggesting efficient analytics for SMEs.
Analytical workloads commonly found in small or medium enterprises are rarely classified as Big Data, as they typically involve data volumes of up to several hundred gigabytes. Data may be distributed across multiple locations, such as local computers, edge devices, or cloud storage. In such scenarios, employing scalable, cloud-optimized data warehouse solutions like Amazon Redshift or Snowflake is not cost-effective. This work presents the implementation of an analytical platform based on the DuckDB embedded database, designed to support moderate-sized workloads by federatively processing queries from various sources. Performance evaluation demonstrates that this system reduces execution time by up to 92.77% compared to MySQL under the same conditions.
No takes yet. Share an insight, caveat, or question.
Ilić et al. (2025) studied this question.