EDB Vice President Franck Sidi: “The Shift to Open Source Is Irreversible” — Offers Solutions to Enterprise AI Transformation Challenges
TECHWORLD ·
✦ AI Summary
Franck Sidi, EDB Vice President, said enterprise AI transformation is accelerating changes in data infrastructure.
He said data fragmentation and rising costs are emerging during AI transformation, and proposed a single integrated data platform and open-source adoption as the solution.
EDB announced a platform including EPAS, WHPG, ClickHouse OLAP, and PGAA, along with plans to release WarehousePG 19 in the first half of 2027.
In an interview on the 3rd, Franck Sidi, EDB Vice President of Global Analytics & AI Sales Engineering, explained how enterprise AI transformation is reshaping data infrastructure, saying that the shift is accelerating changes in data infrastructure. He pointed out that data fragmentation and rising costs are emerging during the AI transformation process.
As a solution, Sidi proposed a single integrated data platform. He also explained that companies are sharply increasing demand for open-source adoption in order to pursue AI transformation, cost reduction, and system innovation at the same time.
He said South Korea is following the same trend, adding that Korean companies are preparing for the next step while using it alongside existing traditional DBs, and that related demand in Korea is also strong. However, he said it was difficult to disclose specific numbers related to companies in market transition, and explained that companies using Oracle and others are rapidly transitioning some systems to Postgres.
The pace of open-source DB adoption differs by industry. While open-source DB usage is active in the gaming and internet business sectors, the transition to open-source DBs in traditional industries such as finance, manufacturing, and the public sector is progressing relatively gradually.
Against this backdrop, EDB offers Oracle-compatible EPAS. The purpose of EPAS is to support migration to Postgres while minimizing code changes to existing systems.
EDB plans to provide similar compatibility features for SQL Server in the future. It also plans to support data migration for MongoDB without directly modifying existing code.
He stressed that the shift to open source offers major cost advantages over commercial DBs for enterprises. EDB said that if companies move to a single Postgres platform, they can reduce costs by about 40% to 50% compared with existing commercial DBs on a TCO basis, including licensing fees. In addition, it said it has completed benchmark tests for WarehousePG on large analytical workloads, and that the cost gap between WarehousePG and Google BigQuery for the same workload is close to 5 times.
The event, 'EDB Postgres AI Summit Seoul 2026,' was held. At the event, the company introduced 'EDB Postgres AI' and the vision for 'Agentic Lakehouse' based on a single SQL and security model, emphasizing their potential to solve enterprise AI transformation challenges.
As AI spreads, the use of DBs and data services by companies is increasing, leading to data fragmentation and cost issues. In response, an integrated platform was presented as a way to improve manageability and reduce costs.
The single integrated data platform consists of WarehousePG (WHPG), ClickHouse OLAP, and Analytics Accelerator (PGAA). WarehousePG (WHPG) was described as being for large-scale DW and analytics, ClickHouse OLAP as for real-time analytics, and Analytics Accelerator (PGAA) as for lakehouse data analysis.
WarehousePG 19 will be officially released in the first half of 2027. The existing Postgres 12-based kernel will be switched to a 19-based one, allowing additional technologies from versions 13 through 18 to be incorporated all at once. According to his explanation, WarehousePG 19 will include Property Graphs, enabling FDS and supply chain analysis using only standard SQL/PGQ without a separate graph DB. He also stressed that WarehousePG 19 will include high-performance Apache Iceberg streaming capabilities.
WHPG is a PB-scale open MPP (Massively Parallel Processing) data warehouse. WHPG supports In-Database ML and ultra-fast analytics, and is positioned as a replacement for Snowflake and Greenplum. According to his explanation, Cloudera can be replaced with WHPG, and there are many customers actively pursuing a shift to WHPG.
ClickHouse OLAP supports sub-second analytics processing for real-time telemetry and log data. PGAA can accelerate queries on Apache Iceberg and Parquet lakehouse data.
PGAA delivers a performance improvement of 3.5 times to up to 5 times compared with existing Postgres. Agentic Lakehouse is based on a single Open Postgres engine.
Agentic Lakehouse integrates transactions, real-time analytics, and AI agents into a single operating environment. He predicted that the shift in the data market from commercial DBs to open-source-based systems will continue.
However, he predicted that a short-term transition of all companies' core systems is unlikely. He said companies will likely continue with gradual transitions while keeping existing systems in place. Ultimately, he said demand will emerge to move even core systems to Postgres.
The speaker explained that while the transition itself is technically possible, the time and effort required are the key variables in practice. Regarding the scope of transition, he said that relatively easy migration methods such as EPAS are possible, and that even shifting core systems to open source is possible, adding that the customer decides which migration approach to choose.
He went on to say that EDB plans to build a platform aligned with these market trends, one that is open, not tied to any specific environment, and provides fully integrated governance. He added that EDB's DNA is data, and that all of its current businesses are 100% focused on the data domain.
Source: TECHWORLD · Kim Hye-jin
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406509
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.