Part II — System Architecture

This section takes you inside the architecture of production vector databases. You'll understand how the core components fit together, how data is stored and replicated, and how hybrid search combines vector similarity with traditional filtering.

Chapters

#	Chapter	Key Topics
6	Core Components	Ingest, index builder, query engine, storage, scheduler, WAL
7	Storage Engines	LSM trees, columnar layouts, delta-merge for mutable vectors
8	Distributed Vector Stores	Sharding, consistency, replication, cloud-native DB comparison
9	Real-Time Update Handling	Dynamic indexes, HNSW layer fan-out, graph-merge
10	Hybrid Search	Score fusion, BM25 + ANN rerank, metadata filtering
11	Hardware Acceleration	SIMD, GPU, FPGA, NUMA pinning, RDMA
12	Observability & Operations	Latency histograms, index health, auto-tuning