Recently, China Academy of Information and Communication Research ("CAICT" for short) officially launched the first batch of vector database product testing of "trusted database". As a representative enterprise of innovation and application in the field of vector database, DingoDB multimode vector database independently developed by DataCanvas participated in and successfully completed this evaluation.

With its rich AI innovation capability and breakthrough AI technology level, DingoDB multimodal vector database completed the test with excellent product capability. It passed a total of 39 test items, including 27 mandatory items. The number of passed items far exceeded the number of evaluation vendors in the same period, becoming the vector database with the most passed items at present.

In order to provide authoritative reference for the supply side R&D and application side model selection vector database, the Institute of CCSA TC601 and CAICT DBL has jointly prepared the Technical Requirements for Vector Database with experts from more than 50 enterprises, It has become the first vector database technology standard in the industry. As one of the core editorial units, DataCanvas provides rich technical reference and cutting-edge practical experience for the formulation of standards.

According to the standard, the basic capability test of vector database products covers 7 capability areas, including basic functions, operation and maintenance management, security, compatibility, scalability, high availability, and tool ecology, with a total of 47 test items, which fully integrates the rich practical experience and wisdom of domestic industry experts, and is the authoritative comprehensive evaluation of the basic capability of vector database. The amazing performance of DingoDB in this test once again confirms that DingoDB's technical level is at the forefront of the industry and will provide a key force for building a new "storage infrastructure" in the era of multimodal large model.

The blue box is required test items, and the purple box is optional test items

As a new paradigm of data processing in the era of big model, DingoDB provides comprehensive data service capabilities of massive storage, multimodal data fusion storage, and joint analysis. It will fully tap the value of data in the era of big model, help users easily complete the two-way empowerment of models and data in the wave of digital intelligence, and bring flexible and efficient data driven decisions and better business development to users.

DingoDB Advantages & Technical Highlights:

● Support mixed structured and unstructured analysis

It provides a hybrid retrieval capability of structured and unstructured indexes, which can simultaneously process and analyze structured and unstructured data, and integrate and comprehensively analyze them, to obtain more comprehensive and accurate analysis results and provide users with a broader data perspective.

● Standard MySQL semantics and powerful vectorization capability

Compatible with MySQL native semantics, users can easily use it, greatly reducing learning costs and migration difficulties; At the same time, built-in embedding function allows users to convert text and image data into vector representation, and perform flexible similarity search and analysis in the database, so as to achieve rapid analysis and retrieval of large-scale text and image data.

● Cluster high availability and massive scalability

Support multi copy storage strategy, effectively respond to the high availability and reliability requirements of data, reduce the risk of data loss, and provide continuously available storage solutions; In addition, DingoDB has good scalability and massive storage capacity, can easily accommodate large-scale datasets, and provides users with flexible resource management capabilities.

● Open-source

Adhering to the product concept of "open-source and openness", it has successfully docked with the large model tool-chain, provided the large model with the ability of massive memory, and ensured the accuracy and reliability of the large model generation results.

● Perfect tool-chain ecology

Through the production level monitoring capability, one click deployment mode and perfect operation and maintenance ecology, it helps users develop, deploy and manage the system flexibly and efficiently, and further improves the work efficiency and system stability.

With the continuous breakthrough of large model technology, vector database will play a more important role in the AI era. This excellent performance in the first batch of vector database product tests of CAICT is an important step that has been encouraged and recognized for DingoDB's new journey towards AI. In the future, DingoDB will continue to innovate and make breakthroughs in technology, providing forward-looking scientific and technological strength for the scenario based landing practice and industrial wide application of vector database technology

DingoDB Official Site：https://www.dingodb.com

DingoDB Github：https://github.com/dingodb