The Challenge
A patent research and risk solutions provider operated a search platform encompassing hundreds of millions of documents. Its patent search architecture needed to support complex document relationships, sophisticated queries, and constant index updates while serving hundreds of clients with customer-specific collections.
The platform’s requirements extended well beyond full text retrieval. Patent content included parent-child relationships and statistical interrelationships, while customer specific annotations added another layer of information to manage and search.
Advanced visual presentations depended on extensive use of search query features. Data modeling, query behavior, and aggregation requirements therefore had to work together to support both retrieval and the ways users explored the information.
The organization also needed to move from an on-premises, single index deployment distributed across multiple shards to a cloud-based, multi-tenant architecture.
Innovent’s Approach
Innovent provided search architecture guidance and implementation support for the transition to a cloud-based, multi-tenant environment. The work addressed how to organize and operate customer specific collections while accommodating substantial data volumes and ongoing updates.
The team advised on ingestion and document modeling, including nested structures and parent-child relationships. This work accounted for the connections within the patent data, statistical interrelationships, and customer specific annotations that needed to coexist with the shared document collection.
Innovent also addressed complex query requirements supporting retrieval and advanced visual presentations. The guidance considered how document structure and query capabilities could work together to expose the information required by the application.
Text analysis and relevance engineering were additional areas of focus, helping align search behavior with the characteristics of patent content and users’ research needs.
Monitoring and cluster design were considered alongside ingestion, modeling, and relevance. This connected the migration effort with the operational demands of a continually updated, multi-tenant search platform.
The Result
Innovent supported the conversion from an on-premises, single index architecture to a cloud-based, multi-tenant solution for a patent research environment encompassing hundreds of millions of documents.
The engagement delivered architectural recommendations and implementation guidance spanning customer-specific collections, complex document relationships, continuous indexing, advanced queries, relevance, and monitoring.
The work demonstrated Innovent’s ability to address information retrieval and infrastructure requirements together, helping evolve a search platform whose data relationships and application demands extended well beyond conventional document search.