Activate Vector Retrieval Service DashVector

更新时间:
复制 MD 格式

Activate DashVector and access the console to start managing your vector data.

Prerequisites

Procedure

  1. Log on to the Alibaba Cloud official website.

  2. Go to the Vector Retrieval Service DashVector product page, and click Activate Now.image.png

  3. Select a Product Type, Region, Instance Type, Specifications, and Number of Replicas. Enter a Cluster Name, and click Buy Now.image

    Parameter

    Description

    Product Type

    The billing method for DashVector. Pay-as-you-go and Serverless are supported. For more information, see Product Billing.

    Region

    The region where DashVector is deployed. The China (Hangzhou), China (Shanghai), China (Beijing), China (Zhangjiakou), and China (Shenzhen) regions are supported.

    Instance Type

    DashVector supports four instance types for different business scenarios:

    • Compute-optimized: Delivers higher QPS and lower query latency. Suitable for high-concurrency, high-traffic scenarios that demand low latency and high write and query efficiency.

    • Storage-optimized: Offers 5 times the storage capacity of compute-optimized instances, allowing you to store and manage more vector data. Suitable for scenarios with large data volumes, rapid data growth, and relatively low QPS.

    • Serverless: Provides unlimited data capacity and automatically scales based on your data. Billing is based on actual requests. Suitable for low-frequency query scenarios (QPS < 2) where latency is not a concern. Overall performance is similar to storage-optimized instances.

    • Free Trial: Designed for testing and evaluation. Do not use it in a production environment. A free trial instance is valid for one month. After it expires, you can apply for another trial. Free trial instances have some limits. For more information, see Limits.

    Note
    • You can create up to 32 collections in a paid cluster.

    • You can create up to 2 collections in a free trial cluster.

    Important
    • Each account can have only one active free trial cluster at a time. After a free trial cluster expires or is manually released, you can create a new one.

    • A free trial cluster is automatically released 30 calendar days after creation, and all data is deleted. If you have important business data, transfer it to a paid cluster or upgrade the free trial cluster to a paid cluster promptly.

    Specifications

    • Free trial cluster: Uses a serverless architecture and is suitable for a quick product evaluation. For the limits of free trial instances, see Limits.

    • Storage-optimized and compute-optimized clusters each offer six specifications, differing mainly in storage capacity. For more information, see Instance specifications.

    • Serverless cluster: Uses a serverless architecture, has no capacity limit, and is suitable for scenarios with low-frequency queries and where latency is not a concern.

    Number of replicas

    You can adjust the number of replicas from 1 to 5. Data is identical across all replicas. More replicas linearly increase QPS and improve service availability. For production environments that require high availability, select 2 or more replicas.

    Note
    • Increasing or decreasing the number of replicas does not affect storage capacity. It only affects QPS and availability.

    • Serverless clusters do not have the concept of replicas.

    Instance name

    Must consist of uppercase letters, lowercase letters, digits, underscores (_), and hyphens (-). The length must be between 3 and 32 characters. An Alibaba Cloud account cannot have two clusters with the same name.

  4. Review the instance information, check the Terms of Service, and then click Activate Now.image.png

  5. After the service is activated, click Management Console to go to the DashVector console.image.png