Case Study: AerData achieves faster, scalable OCR search with Elastic

How AerData improved OCR search capability using Elasticsearch

AerData, a Boeing company, provides STREAM, a document management system that scans, indexes and centralizes technical records for airlines, lessors and MROs. As customer volume and data grew, SQL Server full‑text search became a bottleneck—lacking result ranking, suggestions and easy scale‑out—while complex multi‑filter queries and large backups caused unacceptably slow response times and operational strain, threatening SLAs.

AerData migrated OCR indexing and search to Elasticsearch, using dynamic templates for evolving mappings, a custom tokenizer for partial/full-word and multi‑language search, and the NEST .NET client to build complex queries. They implemented event‑driven synchronization with queued updates and retries, wrapped Elasticsearch with a custom API for security, and automated deployment via Puppet/IaC and CI. The result is a scalable, resilient search platform that handles complex queries, dramatically speeds configuration changes (under 5 minutes in dev/test, ~30 minutes in production), and is easier to maintain and scale.

AerData

Peter Tabangan

Software Engineer