System design interview question
Design a Distributed Search Index and Query System like Elasticsearch at Scale
Design a distributed full-text search platform that indexes and queries massive document corpora in near real-time. The system must support 500M documents, 50K indexing operations per second, and 100K search queries per second across three geographic regions with sub-100ms latency requirements. Documents range from 1KB to 10MB, with complex nested structures and dynamic schema evolution. The system must handle concurrent index updates, support complex boolean/fuzzy/range queries, and provide strong consistency guarantees for recent writes while enabling distributed query execution across shards. Consider how to balance write amplification during indexing, query latency variance across shards, and resource consumption in a cost-sensitive environment.
Practice this question step by step with an AI interviewer that reviews your requirements, your architecture and your trade-offs, and scores the design.