Amazon Cloud Technology vector database preview version now on sale, high performance helps accelerate AI applications

Amazon Cloud Technology launched seven generative AI innovation projects at the New York Summit held on July 26, 2023, which further lowered the threshold for using generative AI, allowing enterprises to focus more on core business and improve production efficiency. One of the eye-catching projects is Amazon Cloud Technology’s upcoming vector database-related innovations. They have released a preview version of Amazon Cloud Technology’s vector engine

Recently, Amazon Cloud Technology released a preview version of the Amazon OpenSearch Serverless vector engine. This release marks a major advancement in cloud search services, providing users with simple, high-performance and scalable similarity search capabilities

In February 2023, Amazon Cloud Technology has been rated as the leader in cloud database management systems by Gartner for eight consecutive years. This honor is not accidental, but a full affirmation of Amazon Cloud Technology's unremitting pursuit of technological innovation and excellence.

So, what is the performance of the preview version of Amazon Cloud Technology's vector engine? Can it afford the public’s expectations of it?

We all know that in this era, generative AI is being rapidly adopted by various industries because it can process big data, automate content generation and provide human-like interactive responses. AI applications such as integrated chatbots, question-and-answer systems, and personalized recommendations use natural language search and query to understand semantics, user intent, and generate anthropomorphic responses, which have revolutionized user experience and digital platform interaction.

Machine learning search and generative AI applications require the use of vector embeddings to represent digital forms of text, images, audio and video to generate dynamic content. These embeddings are trained on user data to express the semantics and context of the information. This process does not need to rely on external data sources or applications. Users hope that the vector database can be easily built and quickly moved from prototype to production environment so that they can focus on differentiated applications

The Amazon OpenSearch Serverless vector engine was launched based on these changes in needs. It extends the search capabilities of Amazon OpenSearch and can store, search, and trace billions of vector embeddings in real time to achieve similarity matching and semantic search. No need to consider infrastructure issues

Therefore, its performance can be roughly summarized as the following characteristics:

The rewritten content is: First, the Amazon OpenSearch Serverless vector engine trial version is naturally robust. Users don’t need to worry about back-end infrastructure selection, optimization, and scaling. The engine automatically adjusts resources to adapt to changing workloads and demands, ensuring fast performance and right scale at all times. Whether the number of vectors increases from thousands to hundreds of millions, the engine can scale seamlessly without reindexing or reloading data, making infrastructure expansion more convenient

Rewritten content: Second, independent computing resources. The vector engine provides independent computing resources for indexing and workload search, enabling seamless acquisition, update, and deletion of vectors in real time, ensuring that user query performance is not affected. Data is stored long-term in Amazon S3 with the same data durability guarantees. Although in preview, the engine is designed for production environments and has redundancy mechanisms to deal with outages and failures

Third, the results provided are accurate and reliable. Customers use OpenSearch kNN search in managed clusters to implement semantic search and personalized recommendations for applications. The vector engine provides the same user experience as the Serverless environment and is simple and easy to use. Amazon OpenSearch serverless vector engine is based on the k-nearest neighbor (kNN) search function of the OpenSearch project. It supports distance indicators such as Euclidean distance, cosine distance, and dot product. It can accommodate 16,000 dimensions and is suitable for various basic models and AI/ML models. Can provide users with accurate and reliable search results

Amazon Cloud Technology plans to launch two features to reduce first-time collection costs for customers. In addition to the great performance mentioned above, first of all, they will launch a new dev-test option that allows users to launch collections without backups or replicas, thereby reducing the cost of entry by 50%. Data durability is still ensured via the vector engine saved in Amazon S3. Secondly, they will also provide an initial phase configuration of 0.5 OCU resources, which can be expanded based on actual workload needs to further reduce costs. This feature works with tens to hundreds of thousands of vectors (depending on the dimensions). In addition, Amazon Cloud Technology has lowered the minimum required OCU from 4 per hour to 1 per hour to provide more support

Amazon Cloud Technology’s ambitions certainly don’t stop there. They are also continuing to work hard to optimize the performance and memory usage of vector graphics, including improving functions such as caching and merging

In the near future, we look forward to Amazon Cloud Technology officially launching the OpenSearch Serverless vector engine. By then, generative AI may enter a whole new field

The above is the detailed content of Amazon Cloud Technology vector database preview version now on sale, high performance helps accelerate AI applications. For more information, please follow other related articles on the PHP Chinese website!