Advertise
AIComputers

UNet unveils TaranTuner database tuning tool at PGConf.Asia 2026

UNet presented TaranTuner, which traces SQL root causes rather than only monitoring, and cited a TV home shopping customer case at the Hong Kong conference.

UNet unveiled its self-developed database performance diagnostics and tuning solution TaranTuner at PGConf.Asia 2026, held in Hong Kong on November 17-18, according to datanet.co.kr.

UNet executive director Lee Kang-il gave a presentation on tracing the SQL root causes of performance degradation beyond basic monitoring. The talk outlined six approaches: ASH collection, SQL ID-based cause tracing, integrated DB/OS analysis, anomaly detection, rapid incident response and before/after performance comparison.

UNet also presented customer cases. A TV home shopping company cut average query response time from 140ms to 32ms and CPU usage from 50% to 25%, while raising TPS from about 1,000 to over 2,000. A retail company case reduced monthly recurring SQL degradation response times from over 2 seconds to about 0.2 seconds.

The company showcased its product lineup alongside TaranTuner: TarantulaDB, Tarantula Lakehouse and Tarantula AI VectorDB. UNet said it aims to expand global partnerships with database firms, cloud providers, SI companies and local IT service providers.

Conference context

PGConf.Asia is described as Asia's leading open-source PostgreSQL technical conference. This year marks the 30th anniversary of PostgreSQL's creation, and the event covers database architecture, cloud native, performance optimization and AI/vector databases.

Quick answers

What did UNet present at PGConf.Asia 2026?

UNet unveiled its TaranTuner database performance diagnostics and tuning solution and showcased TarantulaDB, Tarantula Lakehouse and Tarantula AI VectorDB.

Where and when was PGConf.Asia 2026 held?

The conference was held in Hong Kong on November 17-18, 2026.

What results did UNet report from customer cases?

A TV home shopping company cut average query response time from 140ms to 32ms and CPU usage from 50% to 25%, raising TPS from about 1,000 to over 2,000. A retail company reduced monthly recurring SQL degradation response times from over 2 seconds to about 0.2 seconds.

Source