免费领取大会全套PPT    

点此领取

立即参会

Chunhui Shen

AliCloud Senior Technical Expert

AliCloud multimodal database Lindorm product technical leader, Apache HBase PMC, has more than ten years of practical experience in the field of databases and big data, participated in the development of HBase, HDFS, ElasticSearch, Spark and other systems, and is currently committed to the creation of the next generation of multimodal hyper-converged databases.

Topic

AliCloud Lindorm - Massive Multimodal Data Service for the AI Era

Introduction: Introducing Lindorm, a strategic product of AliCloud, focusing on its technical architecture and conveying the concept of data services in the AI era. In the AI-driven digital era, various types of rich and diverse applications produce a large amount of semi-structured and unstructured data, as well as rich and colorful data management and analysis needs. Traditionally, customers need to choose a different data service for each type of data scenario and connect them through complex data links. As a result, the cost of multiple data stores, the cost of maintaining multiple data services, and the cost of developing on multiple data services pose a huge challenge for agile iterative development of applications.Lindorm is a multimodal database service designed for massive data storage management and analysis in the AI era. , vectors, temporal sequences, etc. It allows users to analyze and manage multiple types of different data internally with a unified set of metadata and SQL access interfaces, and provides full-text search, vector search, AI reasoning and other special analytical functions for semi-structured and unstructured data. For Internet application scenarios, Lindorm can smoothly replace Hbase, providing 100% compatibility, 3x throughput and 2x data compression advantages. For industrial application scenarios, Lindorm provides great simplification of the data platform architecture, making the user's operation and maintenance efficiency and iteration speed increased by more than 2 times, which makes Lindorm popular in the automotive industry, 60% of Chinese automotive companies choose to use Lindorm to build their car networking platform. For AI application scenarios, Lindorm provides one-stop data labeling and vector and full-text fusion retrieval, which can provide 50% cost savings and 2x throughput efficiency compared to similar open source products. Today, more than half of Chinese AI startup unicorns choose Lindorm to build their intelligent search applications. Overall, Lindorm can dramatically improve the cost efficiency, O&M efficiency and development efficiency of application building in the AI era, and help industries quickly realize business value.

© boolan.com 博览 版权所有

沪ICP备15014563号-15

沪公网安备31011502003949号