Global Search vs Federated Search
- Mieko Hashimoto
Global search and federated search might sound similar, and they often do similar things—helping people find information across different places. But they actually work in different ways and are used for different reasons. Understanding the difference between them can help you choose the right approach when building or using a search feature in a system.
1. Scope vs. Source
Aspect | Global Search | Federated Search |
|---|---|---|
Definition | A search feature that allows users to search across all available data and content types within a single platform or system. | A search method that queries multiple distinct, often distributed, data sources (possibly on different systems) and aggregates the results. |
Scope | Typically internal to one platform or ecosystem (e.g., all data in Synapse). | Cross-system or cross-repository search (e.g., searching Synapse, NIH databases, and GitHub simultaneously). |
Goal | Convenience and comprehensiveness within one system. | Access and unify data that’s physically or logically separated. |
2. Architecture
Feature | Global Search | Federated Search |
|---|---|---|
Data indexing | Usually relies on a central index (e.g., ElasticSearch/OpenSearch) where all content is pre-crawled and indexed. | Each source maintains its own index; queries are sent to all sources in real time, and results are combined. |
Performance | Faster, since all data is already indexed. | Slower, because queries are sent across multiple systems on-the-fly. |
Result Ranking | More cohesive and consistent, often uses a single scoring model. | Harder to normalize ranking, since results come from heterogeneous sources. |
3. Use Cases
Global Search
Searching within a large, feature-rich product (e.g., Synapse, Confluence).
Unified access across modules like datasets, users, files, and discussions.
Federated Search
Discovery across platforms (e.g., http://clinicaltrials.gov + PubMed + Synapse).
Search across government, academic, and private data silos.
Often used in federated biomedical infrastructures, library systems, or multi-institutional collaborations.
🧠 Example to Illustrate
Let’s say you're using Synapse:
A global search feature helps you search across projects, data, code, and wikis within Synapse.
A federated search tool might simultaneously query Synapse, PubMed, dbGaP, and Figshare, then return results from all those sources in a single view.
✅ Summary
| Global Search | Federated Search |
|---|---|---|
Platform-bound | ✅ | ❌ |
Cross-system | ❌ | ✅ |
Real-time query | ❌ (usually pre-indexed) | ✅ |
Uniform ranking | ✅ | ❌ |
Complexity | Moderate | High |
Related past discussions
- PORTALS-3596 - Getting issue details... STATUS
- SWC-7195 - Getting issue details... STATUS
- PORTALS-1168 - Getting issue details... STATUS
- DESIGN-853 - Getting issue details... STATUS
- GFP-64 - Getting issue details... STATUS