[Infographic] Highlights | Database New Features in November 2025

1. Launch of AI Node Feature in PolarDB for MySQL

What it is:
PolarDB MySQL now offers dedicated AI Nodes—specialized compute instances co-located with your database—that enable in-database AI inference using pre-trained models (e.g., for fraud detection, sentiment analysis, or predictive maintenance).

Why it matters:

Eliminates data movement: Traditionally, running AI on DB data requires ETL pipelines to export data to external ML systems—a slow, costly, and insecure process. With AI Nodes, inference happens inside the database, using SQL or simple API calls.
Low latency & high security: Sensitive data never leaves the secure DB environment. Ideal for real-time use cases like transaction scoring or dynamic pricing.
Developer simplicity: No need to manage separate AI infrastructure. Just deploy your model (ONNX/TensorFlow/PyTorch) and call it via SQL:

SELECT ai_inference('fraud_model', user_features) FROM transactions;

Who should care:
Enterprises handling sensitive data (finance, healthcare) and developers building real-time AI apps who want to reduce architecture complexity and compliance risk.

2. PolarDB for MySQL Adds Vector Search

What it is:
Native support for vector similarity search using the OpenSearch-compatible PolarSearch engine. Store embeddings (e.g., from text, images) as vectors and run semantic queries directly in SQL.

Why it matters:

Unified stack: No need to sync data to a separate vector DB (like Pinecone or Milvus). Your operational + vector data lives in one system.
High performance: Optimized HNSW indexing delivers millisecond-latency searches even at scale.
Ecosystem compatibility: Works with popular embedding models (e.g., OpenAI, Sentence Transformers) and tools via OpenSearch protocol.
Example use case:

-- Find products visually similar to an image embedding
SELECT product_id FROM catalog 
ORDER BY vec_distance_l2(image_embedding, ?) 
LIMIT 10;

Who should care:
AI/ML teams building RAG (Retrieval-Augmented Generation), recommendation engines, or multimodal search. Startups wanting to avoid managing multiple data systems.

3. ApsaraDB RDS for MySQL Adds Vector Data Processing

What it is:
Full native support for storing, indexing, and computing on high-dimensional vectors (up to 16,383D)—including HNSW indexes, cosine/L2 distance functions, and vector column types.

Why it matters:

Enterprise-grade reliability: Unlike experimental vector plugins, this is a fully supported, production-ready feature in Alibaba Cloud’s battle-tested RDS.
Scalability: Handles massive vector datasets with ACID compliance and point-in-time recovery.
Cost efficiency: Leverage existing RDS infrastructure—no new licenses or cloud services needed.

Key differentiator vs. PolarDB:
While PolarDB targets cloud-native, high-throughput scenarios, RDS for MySQL offers familiarity, stability, and seamless migration for enterprises already invested in RDS.

Who should care:
Large enterprises modernizing legacy apps with AI features without re-architecting their entire data layer. DBAs seeking a low-risk path to vector capabilities.

4. ApsaraDB RDS for SQL Server Launches Global Read-Only Accounts

What it is:
A single database account that grants read-only access to all current and future databases within an RDS SQL Server instance—automatically.

Why it matters:

Solves a chronic pain point: In multi-tenant or microservices architectures, managing per-database read accounts is error-prone and time-consuming. Now, one account = universal read access.
Future-proof: New databases (e.g., from CI/CD deployments) are automatically included—no manual permission updates.
Security best practice: Enforces least-privilege access for analytics, reporting, or audit tools without risking write operations.

Who should care:
Enterprise DBAs, SaaS providers, and compliance officers managing complex SQL Server environments who prioritize operational efficiency and auditability.

202511_en

Community

[Infographic] Highlights | Database New Features in November 2025

1. Launch of AI Node Feature in PolarDB for MySQL

2. PolarDB for MySQL Adds Vector Search

3. ApsaraDB RDS for MySQL Adds Vector Data Processing

4. ApsaraDB RDS for SQL Server Launches Global Read-Only Accounts

Read previous post:

ApsaraDB

You may also like

Comments

ApsaraDB

Related Products

PolarDB for MySQL

PolarDB for PostgreSQL

PolarDB for Xscale

ApsaraDB RDS for MySQL