Azure Cache for Redis Vector Search

Overview

Azure Cache for Redis provides vector embedding and similarity search capabilities, enabling semantic caching and real-time vector search in Azure cloud applications.

Features

Indexing Algorithms:

• HNSW (Hierarchical Navigable Small World)
• FLAT (brute-force for smaller datasets)

Distance Metrics:

• Cosine similarity
• Euclidean distance (L2)
• Inner product

Integration:

• Azure OpenAI
• Azure AI services
• Semantic Kernel
• LangChain

Use Cases

Semantic Caching:

• Cache LLM responses
• Reduce API costs
• Sub-millisecond retrieval

Real-Time Search:

• Product recommendations
• Content similarity
• User profiling

Vector Functions

• EMBED_TEXT_768: 768-dimensional embeddings
• EMBED_TEXT_1024: 1024-dimensional embeddings
• Custom embedding integration

Benefits

• Fully managed service
• High availability (99.9% SLA)
• Azure ecosystem integration
• Enterprise security

Availability

Azure Cache for Redis (Premium/Enterprise tiers)