This topic provides an overview of the billing of OpenSearch Vector Search Edition.
Billing method
The following table describes the billing method that is supported by OpenSearch Vector Search Edition.
Billing method | Description |
Pay-as-you-go | The pay-as-you-go billing method allows you to use resources before you pay for them. The system generates a bill per hour and deducts fees from the balance of your Alibaba Cloud account based on the instance specifications. This billing method is suitable for short-term use. If you no longer need a pay-as-you-go instance, you can release it to reduce costs. |
Instance fee
You are charged for OpenSearch Vector Search Edition based on instances. The billable items related to instances include instance rental, Query Result Searcher (QRS) worker usage, Searcher worker usage, Searcher worker storage, index storage, and data update resources.
Total instance fee = Instance rental fee + QRS worker usage fee + Searcher worker usage fee + Searcher worker storage fee + Index storage fee + Data update resource fee
Instance rental fee = Unit price of instance rental × Usage duration in hours
QRS worker usage fee = Unit price of QRS worker usage × Usage duration in hours × Number of QRS workers
Searcher worker usage fee = Unit price of Searcher worker usage × Usage duration in hours × Number of Searcher workers
Searcher worker storage fee = (Total storage space of a Searcher worker - Free storage space of a Searcher worker) × Unit price of storage space × Storage duration in hours × Number of Searcher workers
Index storage fee = Unit price of index storage × Storage duration in hours × (Size of indexes in GB - 100)
Data update fee = (Total number of resources used for data update - Number of free resources available for data update) × Unit price of data update × Usage duration in hours
Note: If the total size of indexes in an instance is smaller than 100 GB, the index storage is free of charge.
Fees for instance rental, QRS worker usage, and Searcher worker usage
Pay-as-you-go instances
Region | Billable item | Worker family | Available specifications | Unit price (USD) | Unit |
Singapore and Germany (Frankfurt) | Instance rental | - | - | 0.084094 | Per instance-hour |
QRS worker usage | Local SSD | 2 CPU cores and 8 GB of memory | 0.095179 | Per worker-hour | |
4 CPU cores and 16 GB of memory | 0.190359 | Per worker-hour | |||
8 CPU cores and 32 GB of memory | 0.380719 | Per worker-hour | |||
Cloud Disk | 2 CPU cores and 8 GB of memory | 0.082765 | Per worker-hour | ||
4 CPU cores and 16 GB of memory | 0.165531 | Per worker-hour | |||
8 CPU cores and 32 GB of memory | 0.331062 | Per worker-hour | |||
Searcher worker usage | Local SSD | 2 CPU cores and 16 GB of memory | 0.115009 | Per worker-hour | |
4 CPU cores and 32 GB of memory | 0.230018 | Per worker-hour | |||
6 CPU cores and 48 GB of memory | 0.345026 | Per worker-hour | |||
8 CPU cores and 64 GB of memory | 0.460035 | Per worker-hour | |||
12 CPU cores and 96 GB of memory | 0.690053 | Per worker-hour | |||
16 CPU cores and 128 GB of memory | 0.919855 | Per worker-hour | |||
Cloud Disk | 2 CPU cores and 16 GB of memory | 0.101875 | Per worker-hour | ||
4 CPU cores and 32 GB of memory | 0.203750 | Per worker-hour | |||
8 CPU cores and 16 GB of memory | 0.257138 | Per worker-hour | |||
8 CPU cores and 64 GB of memory | 0.407499 | Per worker-hour | |||
12 CPU Cores and 24 GB of Memory | 0.385707 | Per worker-hour | |||
12 CPU Cores and 48 GB of Memory | 0.496593 | Per worker-hour | |||
16 CPU cores and 32 GB of memory | 0.514276 | Per worker-hour | |||
16 CPU cores and 64 GB of memory | 0.662124 | Per worker-hour | |||
16 CPU cores and 128 GB of memory | 0.814998 | Per worker-hour | |||
24 CPU Cores and 48 GB of Memory | 0.771413 | Per worker-hour |
Recommendations on worker families: If your business requires a low queries per second (QPS) but a large amount of indexes, we recommend that you select an instance type that uses local SSDs. If your business requires a high QPS or the indexes can be stored in the memory, we recommend that you select an instance type that uses cloud disks.
Allocation of Searcher workers: After you purchase Searcher workers, you can allocate replicas and data shards for the instance. The number of Searcher workers must be less than the number of Searcher worker replicas multiplied by the number of data shards. You can modify the number of Searcher worker replicas in the Basic Information section on the Instance Details page and modify the number of data shards in the Table Information section on the Table Management page.
Space occupied by the system: Each Searcher worker of the Cloud Disk type has a fixed storage space to be occupied by the system. The fixed storage space occupied varies with the Searcher worker specifications. In a Searcher worker that has 2 CPU cores and 16 GB of memory or 4 CPU cores and 32 GB of memory, 2 GB of storage space is occupied by the system. In a Searcher worker that has 8 CPU cores and 64 GB of memory, 3 GB of storage space is occupied by the system. In a Searcher worker that has 16 CPU cores and 128 GB of memory, 5 GB of storage space is occupied by the system. If you use Searcher workers of the Local SSD type, no fixed storage space is occupied by the system.
Searcher worker storage fee
Pay-as-you-go instances
Region | Billable item | Searcher worker family | Searcher workers specifications | Free quota | Quota range | Step size for purchase | Unit price (USD) | Unit |
Singapore and Germany (Frankfurt) | Searcher worker storage | Local SSD | 2 CPU cores and 16 GB of memory | 100 GB | [100 GB,200 GB] | 50GB | 0.000081 | Per GB-hour |
4 CPU cores and 32 GB of memory | 200 GB | [200 GB,600 GB] | ||||||
6 CPU cores and 48 GB of memory | 300GB | [300GB,700GB] | ||||||
8 CPU cores and 64 GB of memory | 400 GB | [400 GB,800 GB] | ||||||
12 CPU cores and 96 GB of memory | 600GB | [600GB,1000GB] | ||||||
16 CPU cores and 128 GB of memory | 800 GB | [800 GB,1200 GB] | ||||||
Cloud Disk | 2 CPU cores and 16 GB of memory | 0GB | [100 GB,200 GB] | 0.000195 | ||||
4 CPU cores and 32 GB of memory | [100GB,600GB] | |||||||
8 CPU cores and 16 GB of memory | [100GB,200GB] | |||||||
8 CPU cores and 64 GB of memory | [100GB,800GB] | |||||||
12 CPU Cores and 24 GB of Memory | [100GB,500GB] | |||||||
12 CPU Cores and 48 GB of Memory | [100GB,700GB] | |||||||
16 CPU cores and 32 GB of memory | [100GB,600GB] | |||||||
16 CPU cores and 64 GB of memory | [100GB,800GB] | |||||||
16 CPU cores and 128 GB of memory | [100GB,1200GB] | |||||||
24 CPU Cores and 48 GB of Memory | [100GB,700GB] |
Index storage fees
In OpenSearch Vector Search Edition, you are charged for index storage based on the size of the indexes.
If the total size of indexes in an instance is smaller than 100 GB, the index storage is free of charge. If the total size of indexes is greater than 100 GB, you are charged for the excess index size based on the unit price of index storage.
Pay-as-you-go instances
Region | Billable item | Unit price (USD) | Unit |
Singapore and Germany (Frankfurt) | Index storage | 0.000049 | Per GB-hour |
Data update resource fee
OpenSearch Vector Search Edition charges you for data update based on the number of resources that you use.
OpenSearch provides you with a free quota for data update by data source or table in an instance. Each data source or table is provided a free quota of two resources, and each resource consists of 4 CPU cores and 8 GB of memory for data update. You can also apply to increase the quota, and you are charged for the resources that exceed the quota.
Pay-as-you-go instances
Region | Billable item | Unit price (USD) | Unit |
Singapore and Germany (Frankfurt) | Number of resources used for data update | 0.101095 | Per resource-hour |