Write-Read Decoupling in Modern Large-Scale Search Engines: Architectures, Techniques, and Emerging Approaches

2026-05-02

AI Summary

This work addresses severe resource contention between indexing and querying in large-scale search engines, which leads to increased query latency. To mitigate this, the authors propose ScaleSearch, a novel architecture that incorporates five core mechanisms: node-level read-write separation, decoupling of compute and storage, fully in-memory indexing, log-structured write paths, and field-level update routing. These mechanisms collectively enable O(1) in-place updates for scalar fields and isolated processing for full-text fields. ScaleSearch substantially alleviates read-write interference, achieving both low query latency and high index freshness. The architecture thus provides an efficient and real-time system foundation for emerging hybrid retrieval scenarios combining vector and full-text search.

Abstract

Large-scale search engines face a fundamental tension: the index must be updated frequently to maintain freshness, yet updates create resource contention that inflates query latency. In the dominant Lucene-based architecture, segment merges triggered by writes compete with concurrent queries for CPU cycles, disk I/O bandwidth, and operating-system page cache -- a problem we term \emph{write-read contention}. This survey systematically examines the architectural solutions that industry and academia have developed to decouple write pressure from read latency. We identify five principal patterns: (i)~node-level read-write separation; (ii)~compute-storage separation; (iii)~full in-memory indexing; (iv)~log-structured write paths; and (v)~in-place partial updates. We survey representative systems including Elasticsearch, LinkedIn Galene, Uber Sia, Quickwit, Alibaba Havenask, Algolia, Milvus, and Vespa, and discuss an emerging synthesis -- the ScaleSearch architecture -- that combines compute-storage separation with full in-memory indexing and dedicated write nodes. A key contribution of ScaleSearch is \emph{per-field update routing}: each field is assigned its own Kafka topic and update path, allowing scalar fields (price, stock, tags) to be updated in-place in $O(1)$ RAM with immediate visibility while full-text fields follow the segment-based compute-storage path. We conclude with open challenges in hybrid vector-and-full-text retrieval, serverless deployments, and AI-integrated search.

Problem

• write-read contention
• large-scale search engines
• index freshness
• query latency
• resource contention

Innovation

• write-read decoupling
• per-field update routing
• compute-storage separation
• in-memory indexing
• ScaleSearch