📊 HiddenMerit Daily · Issue 34

Focus on Database Frontiers, Practical Insights for DBAs June 1, 2026 ｜ 5 Selected Global Breaking News

01｜Tencent Cloud Database Fully Upgraded to AI‑Native 3.0 Era: Taking Agents as New Users, Reshaping Database Capabilities

On May 29, Tencent Cloud officially announced in Shanghai that its database product system has been fully upgraded for agent scenarios, focusing on three major use cases: agent applications, AI‑assisted programming, and intelligent operations. The upgrade delivers multiple AI‑native database capabilities, marking Tencent Cloud database’s official entry into the AI‑Native 3.0 era.

Wang Yicheng, Vice President of Tencent Cloud, stated at the launch that over more than a decade of development, Tencent Cloud database has successively withstood the dual tests of the internet high‑concurrency wave and autonomous domestic replacement. Now, databases have entered a new AI‑Native 3.0 era. Tencent Cloud has established agents as the new service users, redesigning the database product architecture and capability system from the ground up.

This upgrade brings more than ten technical innovations, including database branching, massive multi‑tenancy, multi‑modal hybrid search, and open storage architecture. Two new benchmark products for agent‑native scenarios were launched:

Agent Memory: Builds three layers of capability: short‑term memory compression, long‑term memory沉淀, and team memory organisation. In long‑task scenarios, it helps agents achieve a 30% increase in task success rate and saves up to 60% in token costs. In the PersonaMem professional evaluation dataset, it raised OpenClaw’s long‑term memory score from 48% to 76%. This product is available with one click in Tencent Cloud’s enterprise‑grade agent platform Claw Pro, and the open‑source version has already gained over 4,000 stars within one week of release.

DatabaseClaw: A database AI agent product built on the OpenClaw framework, embedding the real experience of tens of thousands of DBA tickets at Tencent Cloud, solidifying the troubleshooting SOPs of top DBAs into reusable Skills. A single agent can uniformly cover Tencent Cloud’s 14 database products and 1,600+ OpenAPIs, allowing DBAs to complete cross‑console operational tasks using natural language.

Underlying architecture refactoring: Tencent Cloud refactored the architecture of TDSQL‑B, unifying four engine types – transaction, vector, search, and graph computing – into a single distributed foundation, natively supporting multi‑modal hybrid search. TDSQL‑C introduces database branching capability, reducing the replication time for a 1TB database from hours to seconds. Combined with serverless second‑level startup and idle‑to‑zero capabilities, it precisely adapts to the long‑tail load pattern of AI programming: high‑frequency creation, low‑frequency usage. At the same time, the Hunyuan large model has been integrated into the optimiser, reducing the average latency of “slow SQL” by over 60% .

• DBA Perspective: Tencent Cloud’s repositioning of “agents as new users” represents a paradigm shift in the database industry. Agent Memory’s “team memory” concept means that enterprise operations knowledge can be systematically沉淀 and reused. DatabaseClaw frees DBAs from tedious cross‑console troubleshooting. Most notably, integrating the Hunyuan large model into the optimiser reduces slow SQL latency by 60% – this is not incremental optimisation, but a fundamental reshaping of the database kernel by AI. The DBA role is shifting from “hand‑writing SQL optimisation” to “training and tuning AI optimisation strategies.”

• CTO Perspective: This Tencent Cloud upgrade covers the entire chain from underlying architecture (TDSQL‑B unified multi‑modal foundation), development efficiency (database branching second‑level forking), to operational intelligence (DatabaseClaw). For CTOs planning data architectures for AI applications, this “agent‑native infrastructure” blueprint provided by Tencent Cloud is worth systematic evaluation – especially multi‑modal hybrid search and Agent Memory, which directly solve the most challenging problems for agent applications: long‑term memory and multi‑modal data management.

• Investor Perspective: Tencent Cloud’s intensive investment in the database+AI direction (previously TDSQL OLTP +50%, OLAP +20x), combined with this AI‑Native 3.0 upgrade, is building a full‑chain data intelligence closed loop covering “storage → governance → analysis → agents.” The quality of customer cases disclosed at the launch (Tencent’s internal products such as Yuanbao, QClaw, and CodeBuddy have already been deployed at scale) is a core bellwether for judging commercialisation progress.

Source: Tencent Cloud Official Launch Information

02｜Oracle Releases May 2026 Critical Patch Update: Fixes CVE-2026-46833 and Other Net Service High‑Risk Vulnerabilities

On May 27, Oracle officially released its May 2026 Critical Patch Update (CPU). This update involves vulnerability fixes across multiple product families, with multiple high‑risk vulnerabilities found in the Net Service component of the Oracle Database Server.

Key vulnerability information:

CVE-2026-46833: Located in the Net Service component of Oracle Database Server, affecting versions 23.4.0 through 23.26.2. This vulnerability is “difficult to exploit” (attack complexity high), allowing an unauthenticated attacker to compromise Net Service via network TLS access. Although the vulnerability is in Net Service, the attack may significantly impact other products (scope changed). Successful attack can lead to full takeover of Net Service. CVSS 3.1 Base Score 9.0 (Critical) .

CVE-2026-46834: Also located in the Net Service component, affecting versions 23.4.0 through 23.26.2. This vulnerability is “easy to exploit,” allowing an unauthenticated attacker to compromise Net Service via network TLS access, potentially causing Net Service to hang or repeatedly crash (complete denial of service). CVSS 3.1 Base Score 7.5.

CVE-2026-46835: Similar to CVE-2026-46834, another easily exploitable vulnerability in the Net Service component, leading to denial of service.

• DBA Perspective: The CVSS 9.0 rating of CVE-2026-46833 deserves high vigilance. Although “difficult to exploit” gives DBAs some buffer time, the impact of “no authentication required, network TLS access only, potential full takeover of Net Service” is extremely severe. The affected versions (23.4.0 to 23.26.2) cover most current Oracle 23ai deployments. DBAs should immediately assess the version status of Oracle Database in production environments and schedule a patch window for the May CPU. Since the vulnerability is in Net Service – the core entry point for database network communication – the patch priority should be set to P0.

• CTO Perspective: The multiple vulnerabilities in the Net Service component in this Oracle CPU, especially the “scope changed” characteristic of CVE-2026-46833, indicate that attacks could extend from the database service to other affected products. For enterprises using the Oracle technology stack, this means that a database security vulnerability could lead to cascading risks across the entire application chain. It is recommended to make the May CPU a mandatory item in quarterly security baselines.

• Investor Perspective: Oracle’s sustained high‑frequency CPU releases (consecutive updates in April and May 2026) reflect the ongoing maintenance costs of traditional database security. The 9.0 rating of CVE-2026-46833 also signals that even the most mature commercial databases can have high‑risk flaws in their network service layer. This will continue to drive enterprise customers to evaluate the security and maintenance costs of cloud‑native and open‑source alternative solutions.

Source: Oracle Official Security Advisory, VulDB Analysis

03｜MongoDB Q1 FY2027 Earnings Beat Expectations: Revenue $688M Up 25% YoY, Full‑Year Outlook Raised

On May 29, MongoDB released its first quarter fiscal year 2027 earnings report. Earnings data showed that MongoDB’s Q1 revenue was $687.62 million, up 25% year‑over‑year, exceeding analyst expectations of $667.82 million; adjusted earnings per share were $1.32, also exceeding the expected $1.18.

Other key data: Atlas cloud database revenue grew approximately 27% year‑over‑year, with its share of total revenue continuing to increase. The company held approximately $2.4 billion in cash and equivalents at the end of the quarter. On customer growth, the company added approximately 1,600 new customers, with the total customer count continuing to expand. Based on strong Q1 performance, MongoDB raised its full‑year FY2027 outlook, signalling management’s confidence in future growth.

MongoDB 8.3, positioned as “built for AI speed,” was officially released in early May, delivering, compared to version 8.0, a 35% increase in write throughput, 45% increase in read throughput, and 15% increase in ACID transaction throughput – all without any application code changes. Atlas Vector Search added a public preview of Automated Voyage AI Embeddings, automatically generating embeddings when data is written or updated.

• DBA Perspective: MongoDB 8.3’s performance gains (+35% write, +45% read, +15% transaction) – achieved without code changes – are highly attractive upgrade reasons for DBAs. Atlas Vector Search’s automated embedding generation capability significantly lowers the barrier to building RAG applications on MongoDB. MongoDB’s strong growth also means that DBAs with non‑relational database skills will see their bargaining power in the market continue to increase.

• CTO Perspective: MongoDB’s Q1 earnings beat and raised full‑year outlook, coupled with 27% Atlas cloud revenue growth, prove that its AI‑era platform strategy is gaining market acceptance. The 8.3 release’s positioning as “built for AI speed” and the continued iteration of Atlas Vector Search make MongoDB an important option for AI application data architecture. When evaluating data platforms, CTOs should include MongoDB’s AI capabilities in comparative assessments alongside relational databases.

• Investor Perspective: MongoDB’s Q1 revenue of $688 million, up 25% year‑over‑year, beat market expectations, driving share price increases. The continuing increase in Atlas cloud revenue share (now above 70%) means a higher proportion of sticky, predictable subscription revenue. Raising the full‑year outlook signals management’s confidence in the commercialisation of AI workloads. Investors should continue to monitor customer adoption rates of MongoDB’s vector search capabilities and the revenue contribution proportion.

Source: Benzinga Earnings Report

04｜Kingware Defines New Standard for Real‑Time Data Integration: From “Connecting Data” to “Data as a Stream,” Reshaping Domestic Database Architecture

Recently, CETC Kingware published a technical thought article titled “Kingware Defines New Standard for Real‑Time Data Integration, Reshaping Domestic Database Architecture.” The article points out that when the latency of data flow is compressed from “seconds” to “milliseconds,” we are no longer facing a mere improvement in performance metrics, but a fundamental reconstruction of business logic.

Core judgment: True real‑time capability must be built on seamless collaboration between storage and compute engines. KingbaseES V9, by building a unified metadata management and real‑time synchronisation mechanism, achieves data that can be instantly perceived by multiple endpoints at the moment of writing – data no longer needs to be “moved,” but naturally flows within the system as a “stream.”

In a financial core system scenario, a large bank used the Kingware solution to achieve full‑chain real‑time visibility of transaction data, with query response time stable at under 50 milliseconds. This real‑world data demonstrates that domestic databases are capable of supporting financial‑grade real‑time business. The Kingware solution has passed multiple national security certifications and industry authoritative tests, establishing its reliability in critical infrastructure.

• DBA Perspective: The value of Kingware’s article lies in its core concept – “data no longer needs to be moved, but naturally flows within the system as a stream.” The traditional ETL pattern requires DBAs to repeatedly coordinate data formats, consistency checks, and latency monitoring between sources and targets. By internalising real‑time integration capabilities as a native database feature (rather than external components), Kingware significantly reduces the management complexity of data pipelines. The sub‑50‑millisecond query response time at a large bank is a powerful validation of domestic databases in financial core real‑time scenarios.

• CTO Perspective: The article’s judgment of moving “from ‘connecting data’ to ‘data as a stream’” hits the core pain point of current enterprise data architecture – over 60% of enterprise applications face response latency issues for high‑concurrency real‑time queries due to excessive architectural coupling. CTOs should stop relying on external middleware to patch real‑time capability and instead prioritise evaluating whether the database kernel has native real‑time integration capability.

• Investor Perspective: Kingware, by internalising real‑time integration as a native database capability, marks the transformation of domestic databases from “replacements” to “definers.” In the deep waters of digital transformation, real‑time data integration capability has become a core requirement for enterprise customers. Database vendors with native real‑time integration capability will gain stronger pricing power and order premiums in the next phase of Xinchuang replacement.

Source: CETC Kingware Tech Blog

05｜Kingware Core System Migration Methodology: From “Unthinkable” to “Praktikable,” Dual Validation with Finance and Securities Cases

Recently, CETC Kingware disclosed a series of core system domestic replacement cases. In the migration of a head securities firm’s core trading system, the team conducted over 50 simulation switchover drills, optimising trading response latency from 45ms to 32ms (a 29% improvement), increasing system throughput from 12,000 TPS to 18,000 TPS (a 50% improvement), and compressing recovery time (RTO) from 15 minutes to under 30 seconds (a 97% reduction). In the core system migration of a state‑owned bank, the team achieved a reduction in average core transaction response time from 120ms to 85ms (a 29% improvement), RTO shortened from 4 hours to 15 minutes (a 93% reduction), and single‑node TPS increased from 5,000 to 8,500 (a 70% improvement).

KingbaseES V9 also delivered impressive real‑world migration validation data at a leading news website’s CMS: TPS increased from 12,500 to 16,250 (+30%), storage costs reduced by 48%, full‑text search latency decreased from 450ms to 120ms (-73%), and multimedia read latency decreased from 85ms to 28ms (-67%).

New standard for real‑time data integration capability: Kingware proposes that true real‑time capability must be built on seamless collaboration between storage and compute engines. KingbaseES V9, by building a unified metadata management and real‑time synchronisation mechanism, achieves data that can be instantly perceived by multiple endpoints at the moment of writing.

• DBA Perspective: The RTO data from the securities and banking dual cases (15 minutes → 30 seconds and 4 hours → 15 minutes, respectively) are hard proofs of the high availability capability of domestic databases. When making core system Xinchuang technology selections, DBAs should include similar RTO metrics (financial‑grade requirements are typically within minutes) as a core baseline for evaluating the high availability capability of domestic databases. The article specifically notes that “true ‘zero downtime’ stems from extreme insight into business peak‑valley patterns and the organic combination of phased disaster recovery drills” – this migration methodology proposed by Kingware is worth in‑depth study by DBAs.

• CTO Perspective: The compression of RTO from 15 minutes to under 30 seconds at a securities firm means that the high availability capability of domestic databases in financial core scenarios has reached or exceeded the level of mainstream commercial databases. The data from the state‑owned bank – TPS increased from 5,000 to 8,500 (+70%) and response time reduced from 120ms to 85ms (+29%) – provides financial industry CTOs with highly compelling quantitative selection criteria.

• Investor Perspective: “Zero‑downtime” migration cases at a securities core system and a state‑owned bank core system are the scarcest commercial moats for domestic database vendors. Kingware’s dual‑case validation in the financial sector (bank + securities), combined with the cost‑reduction and efficiency data from the CMS scenario (storage cost -48%), will directly translate into order competitiveness in the deep‑water financial Xinchuang replacement.

Source: CETC Kingware Tech Blog

📅 Recent Database Hot Topics Recap

📌 Issue Summary

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