Amazon Keyspaces (for Apache Cassandra) pricing

Amazon Keyspaces offers two throughput capacity modes for reads and writes: on-demand and provisioned. You can choose your table’s throughput capacity mode to optimize the price of reads and writes based on the predictability and variability of your workload. You also can change the capacity mode of a table once per day as you learn more about the nature of your workload, or if you anticipate having a large burst in traffic, such as from a one-time or seasonal event.

With on-demand capacity mode, you pay for only the reads and writes that your application actually performs. You do not need to specify in advance how many reads and writes you expect your application to perform. Amazon Keyspaces accommodates workloads instantly as they ramp up or down, making it a good option for applications with unpredictable traffic.

Provisioned capacity mode helps you optimize the price of throughput if you have predictable application traffic and can forecast capacity requirements in advance. With provisioned capacity mode, you specify the number of reads and writes per second that you expect your application to perform. You can increase and decrease the provisioned capacity for your table automatically in response to actual application traffic by enabling auto scaling.

Amazon Keyspaces stores data durably across multiple Amazon Web Services Availability Zones using a replication factor of three for high availability.

You can execute reads using either LOCAL_ONE or LOCAL_QUORUM consistency. LOCAL_ONE consistency optimizes for performance and availability by returning the first returned value from any storage replica. LOCAL_QUORUM consistency optimizes for data correctness by requiring at least two storage replicas to return a value before it is returned to your application.

All writes in Amazon Keyspaces are executed with LOCAL_QUORUM consistency for durability.  

Metering of API calls to read and write data in Amazon Keyspaces is based on the capacity mode of the table, the consistency level of the operation, and the size of the data.

 On-demand capacity mode

With on-demand capacity mode, reads are metered using read request units (RRUs) and writes are metered using write request units (WRUs). Each RRU provides enough capacity to read up to 4 KB of data with LOCAL_QUORUM consistency. If you use LOCAL_ONE consistency, you are only billed one half of an RRU for the read. Each WRU provides enough capacity to write up to 1 KB of data per row with LOCAL_QUORUM consistency. For example, to write a row with 3 KB of data requires 3 WRUs. To read a row with 10 KB of data requires 3 RRUs using LOCAL_QUORUM consistency or 1.5 RRUs using LOCAL_ONE consistency. If a query returns multiple rows, you are billed based on the aggregate size of the data returned. For example, if your query returns four rows and each row has 2 KB of data (8 KB of data total), you are billed 2 RRUs using LOCAL_QUORUM consistency and 1 RRU using LOCAL_ONE consistency. 

On-demand capacity mode - Amazon Web Services China (Beijing) Region

You pay a flat price for each read or write request performed:

• Write requests: ¥ 5.575 for every 1,000,000 write request units
• Read requests: ¥ 1.115 for every 1,000,000 read request units

On-demand capacity mode - Amazon Web Services China (Ningxia) Region

You pay a flat price for each read or write request performed:

• Write requests: ¥ 5.575 for every 1,000,000 write request units
• Read requests: ¥ 1.115 for every 1,000,000 read request units

 Provisioned capacity mode

With provisioned capacity mode, reads are metered using read capacity units (RCUs) and writes are metered using write capacity units (WCUs). Each RCU provides enough capacity to read up to 4 KB of data per second with LOCAL_QUORUM consistency. If you use LOCAL_ONE consistency, you are only billed one half of an RCU for the read. Each WCU provides enough capacity to write up to 1 KB of data per row per second with LOCAL_QUORUM consistency. For example, to write one row with 3 KB of data per second requires 3 WCUs. To read a row with 10 KB of data per second requires 3 RCUs using LOCAL_QUORUM consistency or 1.5 RCUs using LOCAL_ONE consistency. If a query returns multiple rows, you are billed based on the aggregate size of the data returned. For example, if your query returns four rows and each row has 2 KB of data (8 KB of data total), you are billed 2 RCUs using LOCAL_QUORUM consistency and 1 RCU using LOCAL_ONE consistency. 

Provisioned capacity mode - Amazon Web Services China (Beijing) Region

You pay a flat, hourly rate based on the capacity you reserve:

• Read throughput: ¥ 0.00116 for every unit of read capacity per hour
• Write throughput: ¥ 0.0058 for every unit of write capacity per hour

Provisioned capacity mode - Amazon Web Services China (Ningxia) Region

You pay a flat, hourly rate based on the capacity you reserve:

• Read throughput: ¥ 0.00116 for every unit of read capacity per hour
• Write throughput: ¥ 0.0058 for every unit of write capacity per hour

With Amazon Keyspaces, you do not need to provision storage in advance. Amazon Keyspaces monitors the billable size of your tables continuously to determine your storage charges.

Amazon Keyspaces offers continuous backups with point-in-time recovery (PITR) to protect your table against accidental writes or deletes. When you enable PITR, Keyspaces backs up your data automatically with per-second granularity, and you can restore your table data to any single second in the preceding 35 days. 

Continuous backups with point-in-time recovery (PITR)

Keyspaces charges for PITR based on the size of each Keyspaces table on which it is enabled. Keyspaces monitors the size of your PITR-enabled tables continuously throughout the month to determine your backup charges and continues to bill you until you disable PITR on each table.

Restoring a table

Restoring a table from continuous backups is charged based on the total size of data restored for each request.

Data transfer “in” and “out” refers to transfer into and out of Amazon Keyspaces. There is no additional charge for data transferred between Amazon Keyspaces and other services within the same Amazon Web Services Region (in other words, ¥0.00 per GB).

Data transfer out rate sizes aggregate outbound data transfer across Amazon DynamoDB, Amazon EBS, Amazon EC2, Amazon Glacier, Amazon Keyspaces, Amazon RDS, Amazon S3, Amazon Storage Gateway, Amazon SNS, Amazon SQS, and Amazon VPC.

Amazon Keyspaces offers Time to Live (TTL), which you can use to set expiration times for table data. Keyspaces TTL pricing is based on the number of expired rows or attributes deleted by TTL. One TTL Deletes unit is consumed per KB of deleted data per row. For example, to delete an attribute in a row that has 2.5 KB of data requires 3 TTL Deletes. If two attributes in the same row expire at the same time, one unit is consumed to delete them (assuming the row is less than 1 KB).

As part of the Amazon Web Services Free Tier, you can get started with Amazon Keyspaces (for Apache Cassandra) for free. If you have never used Amazon Keyspaces before, for the first three months, you are offered a monthly free tier of 30 million write request units, 30 million read request units, and 1 GB of storage per month (limit of one free tier per payer account) in the Amazon Web Services China (Ningxia) Region operated by NWCD. Your free tier starts from the first month when you create your first Amazon Keyspaces (for Apache Cassandra) resource.