< Back to blog home Fujitsu PostgreSQL blog

How Fujitsu helps drive PostgreSQL forward with code-level contributions and features

Marcelo Hauschild | May 7, 2026

 This post focuses on those contributions where impact is most tangible: the features delivered, and the technical foundations laid inside PostgreSQL itself.

A two-decade legacy of open source commitment

Fujitsu's engagement with PostgreSQL dates back to the early 2000s. The company began contributing features as early as PostgreSQL 8.0 and has been recognized as a Contributing Sponsor of the PostgreSQL Global Development Group ever since. That sponsorship is not just a logo on a conference banner, it has funded engineering time, covered travel to events on every continent, and helped ensure that Fujitsu engineers are present wherever PostgreSQL's direction is set.

Fujitsu is also a founding member of the PostgreSQL Enterprise Consortium and a long-standing sponsor of flagship events including PGCon and pgDay. This institutional commitment underpins everything that follows: it creates the conditions for individual engineers to dedicate sustained effort to upstream PostgreSQL, not just to the proprietary features of Fujitsu Enterprise Postgres.

At its core, Fujitsu's philosophy is straightforward: open source innovation is most impactful when organizations participate actively in the communities behind it. The virtuous cycle runs both ways: enterprise-driven requirements inform PostgreSQL development, and upstream advances flow back into Fujitsu's products.

Fujitsu’s contribution impact

Fujitsu’s engineers continued to rank among the top PostgreSQL contributors. The table below shows each engineer's key feature contributions and community participation.

Dominating the frontier of Logical Replication

If there is one technical domain where Fujitsu's leadership is unambiguous, it is logical replication, the backbone of modern high-availability, zero-downtime operations, and multi-master distributed systems. Fujitsu engineers are not just fixing edge-case bugs; they are expanding the conceptual boundary of what PostgreSQL can replicate, when, and how reliably.

Conflict resolution for the Active-Active era

As enterprises move toward distributed, active-active architectures where the same data can be written in multiple locations simultaneously, data conflicts become inevitable. Fujitsu is leading the work on conflict detection and resolution being developed in the community.

• Hou Zhijie co-presented the conflict resolution work at PGConf.dev in Montreal, where the community's core developers assess the direction of future releases. His specialization in the logical replication module has made him the team's de facto architect for parallel apply and row-level filtering.
• Shveta Malik developed a key PostgreSQL 17 feature that enables failover for logical replication slots, ensuring continuity of logical replication after a primary failover or planned switchover. She also contributed to the initial research, design, and review of several other critical features, including PostgreSQL 18 Logical Replication Conflict Detection, PostgreSQL 19 Skipping Tables when Publishing All Tables, and Sequence Synchronization, highlighting deep expertise in designing reliable and correct logical replication systems.
• Ajin Cherian brings the longer view: he introduced two-phase commit support in PostgreSQL 15, ensuring that distributed transactions across subscriber nodes maintain the same ACID guarantees as local ones. He is currently leading work on automatic sequence replication, closing one of the last remaining gaps for teams replicating complex schemas.

Zero-downtime operations and subscriber upgrades

For 24/7 enterprise environments, the question is never can we upgrade? but can we upgrade without stopping? Fujitsu has built the answer into PostgreSQL itself.

• Vignesh C developed the landmark PostgreSQL 17 feature that preserves logical replication states during subscriber upgrades.
  Before this, upgrading a subscriber required a full re-sync, potentially re-transferring terabytes of data. Vignesh eliminated that risk, making maintenance windows far less disruptive. He also served as a Commitfest Manager for one of the Commitfests for PostgreSQL 17, reviewing and shepherding patches from the wider community.
• Shlok Kyal has taken this further with enhancements to pg_createsubscriber that allow organizations to convert existing physical disaster-recovery standby nodes into active logical replicas for reporting, flipping a node's purpose without re-syncing its data over the wire.
• Shubham Khanna continues to refine pg_createsubscriber's usability and flexibility, focusing on the operational patterns required by complex multi-tier deployments. He expanded on the topic in our blog post What’s new in pg_createsubscriber: A look at the latest enhancements.
• Hayato Kuroda contributed to pg_createsubscriber in PostgreSQL 17 and presented the feature at PGConf.dev 2024. He also committed a change that allowed pg_upgrade to migrate replication slots and was officially certified as a PostgreSQL Contributor in January 2025, a formal recognition from the global project of the quality and depth of his work.

Row filtering, tablesync, and generated columns

Not every system needs to replicate every row. Peter Smith has made PostgreSQL's logical replication dramatically more selective and efficient.

His contributions include logical replication row filters (allowing publishers to specify which rows flow to which subscribers), tablesync workers (improving the reliability of initial table synchronization), and replication of stored generated columns, a feature that eliminates the need for subscribers to re-compute derived values locally. Peter also serves as a primary patch reviewer, bringing his decades of C expertise to the quality control of the entire team's output.

PostgreSQL meets the AI era

As the industry pivots toward Artificial Intelligence, Fujitsu is leading the conversation on how the database must evolve to serve it safely. Gary Evans, Senior Offerings and Center of Excellence Manager, argues a foundational truth through his PostgreSQL AI series: enterprise AI starts with the database, not the data swamp.

Fujitsu's vision centers on Retrieval-Augmented Generation (RAG) integrated with Fujitsu Enterprise Postgres. By combining vector search, via the pgvector extension and PostgreSQL's native vector capabilities, with the database's existing ACID guarantees, Fujitsu enables organizations to build AI applications that reason from structured, governed, trustworthy data rather than unmanaged data lakes.

This matters because the risks of AI are largely data risks. An LLM querying a vector index backed by a properly governed PostgreSQL schema inherits all of that schema's access controls, audit logging, and transactional integrity. Fujitsu is building and documenting exactly that architecture, presenting it at conferences and publishing it in its PostgreSQL Zone technical series.

At PGConf.dev, Gary Evans also explored the future of learned indexes, data structures that replace traditional B-trees with machine learning models trained on the data distribution. This is exploratory work at the frontier of database research, and Fujitsu is ensuring PostgreSQL's community is engaged with it early.

Security, compliance, and the enterprise foundation

Enterprise adoption of PostgreSQL depends as much on security and compliance as on performance. Fujitsu addresses this on two fronts.

Gary Evans has written a detailed technical exposition in site blog series Database security, in which he maps PostgreSQL's security capabilities to the NIST Cybersecurity Framework, giving security-conscious organizations a structured path from framework requirement to database configuration.

Rajni Baliyan, Database Support Manager and President of PostgreSQL Down Under (PGDU) committee, focuses on the practical implementation layer, operational excellence, customer success and PostgreSQL adoption globally. Her work on applying CIS Benchmarks to PostgreSQL deployments gives DBAs a hardened, auditable configuration baseline. She is an active global speaker, regularly presenting at international conferences including PGConf APAC, PGDay Australia, FOSSASIA and share insights on PostgreSQL at scale, enterprise adoption, and the evolving compliance landscape for Australian companies. Her blog posts on logical replication and migration from legacy systems have been widely referenced in the Asia-Pacific region. 

Together, these contributions address a common barrier to PostgreSQL adoption in regulated industries: the perception that an open-source database requires more security effort than a proprietary one. Fujitsu's frameworks and documentation demonstrate the opposite.

Why Fujitsu's contribution matters for your business

When you choose Fujitsu Enterprise Postgres, you are not just buying a product. You are partnering with the people who write the code that powers the global PostgreSQL ecosystem.

Fujitsu's upstream contributions mean that the features your DBAs rely on, such as conflict resolution, zero-downtime upgrades, row-level filtering, failover slots, were designed with enterprise requirements in mind from the start. They did not emerge from a research lab; they emerged from production experience at large-scale deployments, channeled back into the community through the pgsql-hackers mailing list and the PostgreSQL commitfest process.

This also means that Fujitsu's engineers understand the codebase at a depth that is genuinely rare. When an issue requires diagnosis at the logical replication layer, the people best positioned to help are the people who wrote that layer.

For organizations evaluating PostgreSQL as a cost-effective, sovereign alternative to expensive proprietary systems like Oracle, Fujitsu's investment in the community provides a level of confidence that is difficult to replicate with a database vendor whose engineers are primarily consumers of upstream work rather than contributors to it.

From upstream code to global community

Taken together, these contributions reflect more than isolated patches or one off enhancements. They represent sustained stewardship of PostgreSQL’s most demanding subsystems. Whether enabling active-active replication, preserving replication state across upgrades, or extending the database to support AI era workloads securely, Fujitsu’s engineers are working at the level where PostgreSQL’s future is decided.

For organizations that depend on PostgreSQL in production, that depth of upstream involvement translates directly into resilience, insight, and confidence, because the people supporting the system are often the same people who built it.

This article has focused on Fujitsu’s upstream PostgreSQL contributions, but we also invest heavily in sharing this expertise through openly accessible resources and direct engagement with the global PostgreSQL community through technical publications, conference participation, and community engagement. That broader dimension is the topic of my next blog post, so be sure to subscribe to new blog posts and stay tuned.

 

 

Ready to run PostgreSQL at enterprise scale?

The engineers who help build PostgreSQL can help you deploy, optimize, and support it.

Explore Fujitsu's PostgreSQL consulting services and put two decades of upstream expertise to work for your organization.

Explore our PostgreSQL Consulting Services >