Unnnet is stepping up its push into the database platform market aimed at the AI market.

Unnnet provides TarantulaDB, a data platform that structurally separates operational and analytics and AI workloads while keeping them organically connected.

The company said TarantulaDB is an enterprise database based on PostgreSQL. It integrates high availability, automatic failover, backup and recovery, security, SQL tuning and heterogeneous integration features. Transaction-focused core operations are handled in this area.

Large-scale analytics and AI use are handled in the Tarantula lakehouse structure. The platform combines the analytics performance of a data warehouse with the scalability of a data lake. It manages structured and unstructured data in a single architecture and links to AI and BI tools.

With this, Unnnet implemented a structure that can run large-scale data analytics and machine learning training in parallel without burdening the operational database.

Unnnet CEO Ju-seob Kim (김주섭) said, "Many companies expect that adopting a high-performance database will naturally expand to AI, but in reality they force analytics and AI workloads onto the operational database, causing performance degradation, rising costs and operational risks at the same time." He added, "For AI, data structure and operating strategy must be designed first, before the model."