Segment 1: HiddenMerit Morning Post · Issue 103

Focus on Database Frontiers, Practical Insights for DBAs September 14, 2026 ｜ 5 Selected Global Breaking News

01｜Dameng Granted New Data Synchronisation Patent: Initial Load and Incremental Sync in Parallel, Significantly Improving Sync Efficiency

Dameng has been granted a new invention patent titled “A Data Synchronisation Method and Device.” The patent divides tables in the source database into multiple groups for initial loading separately. When a target group completes initial loading, incremental synchronisation is performed on the tables in that group — initial loading and incremental synchronisation proceed in parallel, saving incremental synchronisation waiting time and optimising data sync performance. Dameng has received 30 new patent grants this year, up 50% year‑on‑year. H1 R&D investment was RMB 161 million, up 43.9% year‑on‑year.

DBA Perspective: Data synchronisation is one of the most time‑consuming steps in Xinchuang migration. The patent’s “parallel loading + incremental sync” approach directly addresses the industry pain point of “short downtime windows.” For DBAs, this means Dameng’s technical reserves in migration toolchains are accelerating, worth including in Xinchuang project selection evaluations.

📌 HiddenMerit In‑Depth Preview: Today’s in‑depth article – “DuckDB Acquired by Amazon: The ‘Coming of Age’ of a Lightweight Analytics Engine and DBA Architecture Choices” – will explore the industry logic behind this deal.

02｜Amazon Acquires DuckDB Developer DuckLabs: An Analytics Engine Downloaded a Million Times a Day “Comes of Age”

Amazon has announced a definitive agreement to acquire Amsterdam‑based DuckLabs, the developer of the open‑source analytics database DuckDB. DuckDB is downloaded over 1 million times globally each day and is regarded by developers as synonymous with “lightweight analytics” — no need to build an expensive cloud data warehouse; you can query large datasets quickly on your local machine.

Key Deal Terms: Amazon repeatedly emphasised that “we’re buying DuckLabs the company, not DuckDB the open‑source project.” DuckDB itself will continue to be free and open source under the MIT licence, overseen by an independent non‑profit, the “DuckDB Foundation.” The two founders, Hannes Mühleisen and Mark Raasveldt, will join AWS with their 30+ person team, continuing to lead technical direction from Amsterdam.

Strategic Logic: What Amazon is buying into is the trend that “data no longer sits obediently in cloud data warehouses, but is spread directly across cloud storage and queried in place.” This trend has been dramatically amplified by AI — AI agents query data differently from humans; they like to grab a small handful first, try things out, and iterate repeatedly. A lightweight, fast‑starting engine like DuckDB fits this pattern perfectly.

Industry Reaction: MotherDuck, a startup building cloud services on DuckDB, immediately announced it would begin offering enterprise‑grade support for DuckDB, taking a坦然 stance: “We welcome the competition.”

DBA Perspective: DuckDB’s acquisition is a landmark event marking the transition of “embedded analytics engines” from developer tools to cloud infrastructure. For DBAs, this means the architectural choices for data analytics are expanding from “centralised data warehouses” to “lightweight local engines” — especially as AI agents become data consumers, engines like DuckDB with fast startup and low resource usage may become an important complement to agent data access.

03｜Bank of Chengdu Procures Middleware Distributed Database: 20 Licences, Including Three Years of OEM Maintenance

The tender results for Bank of Chengdu’s middleware distributed database procurement project have been announced. The procurement covers 1 set of middleware distributed database software, 20 licences in total, including data migration, management, operations, analytics, monitoring management platform and all product features, along with implementation services and three years of 7×24 OEM standard maintenance service. The first winning candidate is Sichuan Changhong Jiahua Information Products Co., Ltd.

DBA Perspective: The city commercial bank procuring “middleware distributed database” rather than simply database software shows Xinchuang projects are moving from “single‑point replacement” to “full‑stack collaboration.” The procurement scale of 20 licences plus three years of maintenance also reflects increasing investment by city commercial banks in Xinchuang replacement of core systems.

04｜CVE-2026-90514: School Registration and Fee System SQL Injection, Unauthenticated Attackers Can Fully Control Database

SourceCodester School Registration and Fee System 1.0 has been found to have a remote SQL injection vulnerability (CVE-2026-90514, CVSS 7.3) in the Status parameter of the /bilal/normal/save_stud.php script, allowing arbitrary database queries without sufficient filtering. Remote unauthenticated attackers can steal student and staff personally identifiable information (PII), financial records, and even delete database contents.

DBA Perspective: SQL injection vulnerabilities in educational web applications remain a high‑frequency risk. This application is often deployed in school environments to manage student registration and fees; once breached, student privacy data is directly exposed. DBAs in the education sector are advised to work with security teams to identify exposed instances, deploy WAF rules to intercept SQL keywords in the Status parameter, and review web access logs for anomalous POST requests.

05｜Oracle CEO Larry Ellison Fades from Earnings Calls: AI Transformation Takes the Stage, Co‑CEOs Lead Cloud Narrative

Bloomberg reports that Oracle founder Larry Ellison has been absent from the company’s earnings calls this year and is not listed as a keynote speaker at next month’s user conference. Last September, Oracle promoted Clay Magouyrk and Mike Sicilia to co‑CEOs. Magouyrk joined Oracle in 2014 (the same year Ellison stepped down as CEO), previously worked at Amazon and AWS for six years, and single‑handedly expanded Oracle Cloud Infrastructure to over 100 public regions.

Background: Oracle is cutting $2.8 billion in costs while borrowing to build data centres; headcount has fallen 13% in a year. In Europe, the “sovereign cloud” products led by Magouyrk (Frankfurt and Madrid regions operated by separate EU legal entities) have become a key narrative distinguishing the post‑Ellison era.

DBA Perspective: Ellison’s fade from earnings calls is a sign of Oracle’s transition from “founder narrative” to “cloud operations narrative.” For DBAs, this means Oracle’s product roadmap will focus more on cloud infrastructure and sovereign compliance — the priority of traditional on‑premises databases may further decline.

💡 SQL Little Knowledge Point

This Issue’s Knowledge Point: What is DuckDB?

DuckDB is an open‑source embedded analytics database, described by developers as “SQLite for data analytics.”

Dimension	DuckDB	Traditional Cloud Data Warehouse
Deployment	Embedded, runs locally	Centralised, cloud‑deployed
Startup Speed	Extremely fast, seconds	Requires cluster startup time
Resource Usage	Lightweight	High resource requirements
Use Cases	Lightweight analytics, AI agent data exploration	Large‑scale enterprise analytics

Why AI Agents Like DuckDB: AI agents query data differently from humans — they like to “grab a small handful first, try things out, and iterate repeatedly.” Traditional data warehouses’ startup costs and query latency are too heavy for agents, while DuckDB’s lightweight, fast‑starting characteristics are perfectly suited to agents’ exploratory data access patterns.

Implications for DBAs: When AI agents become data consumers, DBAs need to consider not just “can the data warehouse handle high concurrency,” but also “can the data engine be quickly started and called at low cost by agents.” DuckDB’s rise is an early signal of this trend.

HiddenMerit Team Production Slogan: 绩优隐于内，金石启新程 | Hidden deep. Merit bold. Forge ahead.

Segment 2: HiddenMerit In‑Depth · Issue 16

Focus on Database Frontiers, In‑Depth Analysis September 14, 2026 ｜ 1 In‑Depth Feature Article

DuckDB Acquired by Amazon: The “Coming of Age” of a Lightweight Analytics Engine and DBA Architecture Choices

1. A “Buy the People, Not the Project” Acquisition

Amazon has announced a definitive agreement to acquire Amsterdam‑based DuckLabs, the developer of the open‑source analytics database DuckDB.

The most intriguing aspect of this deal is: Amazon is buying the company, not the project.

DuckDB itself will continue to be free and open source under the MIT licence, overseen by an independent non‑profit, the “DuckDB Foundation,” not controlled by Amazon. The two founders, Hannes Mühleisen and Mark Raasveldt, will join AWS with their 30+ person team, continuing to lead technical direction from Amsterdam.

This is not a conventional “absorbing an open‑source project” move, but a carefully designed arrangement of “bringing the core team in, leaving the open‑source project outside.” What Amazon wants is not to monopolise DuckDB, but to make DuckDB an industry standard — once the standard is established, the compute running on top of it will flow continuously to AWS.

2. What Did DuckDB Get Right?

DuckDB’s positioning can be summarised in one sentence: “SQLite for data analytics.”

Traditional analytical databases (such as Snowflake, Redshift) require building clusters, configuring resources, and waiting for startup. DuckDB does the opposite: embedded, runs locally, starts in seconds, extremely low resource usage. Developers can run it directly on a laptop or embed it in CI/CD pipelines.

Global downloads exceed 1 million per day — this figure shows DuckDB has penetrated developers’ daily workflows.

More critically, DuckDB caught the convergence of two trends:

Trend 1: Deepening storage‑compute separation. Data no longer sits obediently in dedicated cloud data warehouses waiting to be moved; it’s spread directly across cloud storage (such as S3), queried in place by whoever needs it. DuckDB’s lightweight engine fits this “storage outside, compute inside” model perfectly.

Trend 2: AI agents becoming the new data consumers. AI agents query data differently from humans — they like to “grab a small handful first, try things out, and iterate repeatedly.” Traditional data warehouses’ cluster startup costs and query latency are too heavy for agents, while DuckDB’s fast startup and low resource usage are naturally suited to agents’ exploratory data access.

[quads id="805"]

An AWS vice president put it bluntly: “DuckDB was born for this kind of AI agent usage.”

3. Amazon’s Strategic Chess Game

Amazon’s calculation is clear: make DuckDB the default analytics engine on S3.

AWS and DuckLabs have been collaborating for two years; DuckDB is already used in Amazon’s own S3 Tables and SageMaker Lakehouse. This acquisition is more of a “coming of age” than a pursuit from scratch.

After acquiring the DuckLabs team, Amazon can:

Deeply optimise DuckDB‑S3 synergy — making the “data in S3, compute in DuckDB” experience seamless
Ensure DuckDB’s technical direction aligns with AWS strategy — such as prioritising support for new S3 features
Attract ecosystem through open‑source standards — when DuckDB becomes “the default analytics engine on S3,” all users doing analytics on S3 will naturally flow to AWS

And DuckDB remaining open source ensures ecosystem vitality and developer trust.

4. Architectural Implications for DBAs

Implication 1: Data analytics architecture choices are expanding from “centralised” to “lightweight”

In the past, a DBA’s first reaction to analytics needs was “build a data warehouse.” But DuckDB’s rise shows that not all analytics requires a data warehouse. Lightweight, embedded, locally‑run analytics engines have irreplaceable advantages over data warehouses in data exploration, CI/CD testing, and edge computing scenarios.

When planning data architecture, DBAs can consider a hybrid model of “data warehouse + lightweight engine”: data warehouses handle large‑scale, high‑concurrency enterprise analytics, while DuckDB handles exploratory, lightweight, agent‑driven data access.

Implication 2: AI agent data access patterns are changing engine selection criteria

Traditional database selection looks at “can it handle high concurrency, can it guarantee transaction consistency.” But AI agents as new users look at “can it start fast, can it iterate at low cost, can it be embedded in agent workflows.”

DuckDB’s acquisition by Amazon is an early signal of “agent‑friendliness” becoming a new criterion in database selection. DBAs need to start understanding agent data access patterns — high‑frequency, small‑batch, exploratory, iterative — and assess whether existing data architectures fit these patterns.

Implication 3: Open‑source project “governance structure” is becoming a strategic asset

DuckDB’s “company acquired, project stays open source” model may become a new paradigm for open‑source database commercialisation. The core team gains commercial returns, the open‑source project maintains independent governance, and the cloud vendor gains standard‑setting power.

For DBAs, this means that when evaluating open‑source databases, governance structure (who manages it, how, what’s the commercialisation path) will be as important as technical capability. An open‑source project with a clear governance structure is more trustworthy than one where “you don’t know who it’ll be sold to in the future.”

5. The “Coming of Age” Moment for Lightweight Engines

DuckDB’s acquisition by Amazon marks a broader trend: lightweight, embedded, AI‑friendly data engines are moving from “developer toys” to “cloud infrastructure.”

When AI agents become the primary consumers of data, when data moves from “centralised storage” to “available anywhere,” when analytics shifts from “batch scheduling” to “instant exploration” — the “coming of age” of engines like DuckDB may just be the beginning.

For DBAs, this is not a threat but an expansion of the architectural toolbox. When you have more than just a “data warehouse” hammer in your toolbox, you can solve more problems.