Migrating from the Google Search Appliance (GSA) to Solr
Reducing costs and improving search for intranet and customer support portal
The customer is a global leader in mobile communications. With the large amount of data the company gathers and publishes, they were looking to migrate from the Google Search Appliance to Solr, an open source search engine, in order to scale in a cost-effective way.
Search Technologies was selected to plan and implement the migration for their two mission-critical search applications:
- Corporate wide search: an intranet search application used by the company’s internal staff.
- Public-facing site search: a support portal search application designed for external users, including customers, to search across documentation and manuals.
The Google Search Appliance is a widely-used search engine that can handle complex enterprise search needs. So migrating from the GSA was a strategic move that required thorough planning for a fail-proof implementation. Aside from a smooth migration to Solr, the customer also wanted to ensure their specific requirements would be met.
The intranet search application needed to be capable of:
- Supporting crawls for sitemap and over eight different secured content sources
- Providing a centralized, secure access with Single Sign On (SSO)
- Handling the same document-level security provided by the GSA
For their support portal, the customer required the search application to:
- Support the crawl of a large list of URLs while still maintaining reasonable performance.
- Have a query completion feature generated from the most popular search queries.
We developed a comprehensive migration plan for implementing and configuring Solr to meet the customer's requirements.
Intranet search application
Since the GSA came with security manager, we implemented authentication and integrated it with a network authentication protocol to support SSO.
We then built a custom user interface (UI) that was integrated with authentication.
Public support portal search application
We divided the large list of URLs into several content sources and leveraged Aspire, Search Technologies’ content processing framework for structured and unstructured data, for efficient crawls and content processing.
For a high-performance query completion feature, the Query Processing Language, a part of the Aspire framework, was used for query parsing, exception handling, and building a custom dictionary from the application’s query logs.
Enhanced search features for both applications
We also implemented the following search features to help users find content quicker and easier:
- Dynamic search navigators
- Spell check
- Search term highlight
- Type-ahead or query suggestions based on popular searches
We successfully migrated millions of documents from the Google Search Appliance to Solr, providing both internal and external users an undisrupted search experience and fulfilling all requirements.
SSO gave internal users easy access while document-level security delivered the right content to the designated user groups (with different access permissions).
The Aspire dashboard helped the customer effectively manage and monitor crawls for various content sources. Content processing via Aspire also allowed for better speed and relevancy – two critical search performance criteria expected by end-users.
If you would like to learn more about the solution architecture and project details, please contact us.