Overview
OpenSearch serves as a one-stop platform used to develop commercial intelligent search services. It is built based on a large-scale distributed search engine platform developed by Alibaba. OpenSearch provides mid-end services for the core search business of Alibaba Group, including Taobao, Tmall, and Cainiao. Thanks to years of search experience in various industries and the capabilities of handling traffic peaks during Double 11, the team of OpenSearch releases a set of service editions that offers high performance, efficiency, availability, and stability. The released service editions are LLM-based Conversational Search Edition, Industry Algorithm Edition, High-performance Search Edition, Vector Search Edition, and Retrieval Engine Edition. This helps meet search requirements in a variety of industries. OpenSearch is a PaaS that simplifies the use of search technologies and lowers the technical threshold and costs. OpenSearch allows you to implement search features and fast iterations for your services at a low cost. This prevents search technologies from being your business bottleneck.
To create a search service by using OpenSearch, you need to only perform the following steps:
Create an OpenSearch instance.
Modify the instance.
Synchronize or upload documents.
Send a search request from an application or web client, access OpenSearch by using API, and obtain search results.
Architecture
LLM-based Conversational Search Edition
Software as a service (SaaS) platform: You can interact with the system by using the OpenSearch console or by calling API operations.
Large language model (LLM) and leading natural language processing (NLP) model: To quickly create a large conversational search model dedicated to an enterprise, you need to only import business data.
Conversational search mode: LLM-based Conversational Search Edition returns search results in conversational mode based on business data. This ensures the security, stability, quality, and reliability of the search results.
Industry Algorithm Edition
SaaS platform: You can interact with the system by using the OpenSearch console or by calling API operations.
Scenario-based configuration: You need to only create an application and configure the data source, fields, and search attributes. After reindexing is complete, you can test the search performance by using an SDK or in the OpenSearch console.
Industry template: OpenSearch Industry Algorithm Edition provides a variety of industry algorithm packages for the e-commerce, content communities, games, and education industry to help you build industry-specific search services. After you specify the industry, you can create search-related features and services, such as semantic understanding, sorting, and search guidance. This helps implement fast access and ensure intelligent industry-specific search results.
High-performance Search Edition
SaaS platform: You can interact with the system by using the OpenSearch console or by calling API operations.
Scenario-based configuration: You need to only create an application and configure the data source, fields, and search attributes. After reindexing is complete, you can test the search performance by using an SDK or in the OpenSearch console.
Big data search: Compared with Industry Algorithm Edition, High-performance Search Edition eliminates the complex industry algorithm capabilities and focuses on the throughput of business queries and writes based on the support for common search capabilities, such as analyzers and sorting. High-performance Search Edition provides the capabilities to respond within seconds and query in real time in scenarios where queries are performed for large data sets.
Vector Search Edition
PaaS platform: You can perform O&M operations on Vector Search Edition instances by using a visual management platform.
Scenario-based configuration: You need to only create an application and configure the cluster, data source, and index table. After reindexing is complete, you can test the search performance by using an SDK or in the OpenSearch console.
Vector search: Vector Search Edition provides you with a variety of query parameters and combined search modes based on inverted indexes and vector indexes. Vector Search Edition is highly reliable and offers high performance in scenarios such as graph searches, audio or video searches, recommendations, and Q&A.
Retrieval Engine Edition
PaaS platform: You can perform O&M operations on Retrieval Engine Edition instances by using a visual management platform.
Scenario-based configuration: You need to only create an application and configure the cluster, data source, and index table. After reindexing is complete, you can test the search performance by using an SDK or in the OpenSearch console.
SQL query: SQL query syntax is supported to reduce your learning costs.
Minimized costs: Retrieval Engine Edition provides various built-in policies for index compression to reduce your business costs and meet your query requirements at lower costs.