NoSQL Database Cloud Service Pricing

Metric description

Per the service description (PDF), the metrics are defined as below:

• Write Unit Per Month: Defined as the throughput of up to one kilobyte (KB) of data per second for a write operation over one month, or approximately 2.6 million writes. Each month is deemed to have 744 hours.
• Read Unit Per Month: Defined as the throughput of up to 1 KB of data per second for an eventually consistent read operation (i.e., where the data returned may not be the most recently written data to the database) over one month, or approximately 2.6 million reads. Each month is deemed to 744 hours. To achieve the throughput of up to 1 KB of data per second for an absolute consistent read operation (i.e., where the data returned is the most recently written data to the database), the equivalent of two Read Units Per Month needs to be provisioned.
• Gigabyte (GB) Storage Capacity Per Month: Defined as 1 GB of computer storage space used by a storage filer of the Oracle Cloud service during a month of the service.

One Write or Read Unit Per Month measures the total consumption of 2,678,400 KB of writes or reads per month (744 hours).

Since there are 2,678,400 seconds in a month:

• If your application writes or reads 1 KB of data every second consistently across 744 hours, it will consume 1 write unit or read unit per month.
• If your application writes or reads 2 KB of data every 2 seconds consistently across 744 hours, it will still consume 1 write unit or read unit per month.

Your application workload characteristics may vary dynamically and consume a different number of write or read units per month; you will be billed accordingly per your Oracle NoSQL Database Cloud consumption.

In addition to writes and reads, storage is another resource required. Storage capacity is provisioned in advance. Capacity can vary within a month; it is measured and billed based on the number of GB capacity provisioned.

Billing examples

Consider an application that performs some write and read operations in different periods of a month and consumes the following write and read:

• Days 1–5: Consumed 100,000 writes and reads per day
• Days 6–15: Consumed 250,000 writes and reads per day
• Days 16–31: Consumed 45,000 writes and reads per day

Assumptions

Assume the following for simplicity:

• The record size for the write or read operation is 1 KB or less.
• Reads are eventually consistent which are billed at 1 Read Unit per 1 KB operation.
• Writes are billed at 1 Write Unit per 1 KB operation.
• No secondary indexing. Records are accessed through primary key(s).
• A month has 31 days and 744 hours.
• Storage provisioned on Day One stays constant at 5 GB throughout the month.

Monthly billing examples

Here's a sample of the monthly billing, assuming both workloads result in the same total amount of write and read units consumed for the 31 days.

Billing example for both workloads in a month (31 days)

Writes
Workload period	Daily write workload	Total writes consumed for each period	Details
Day 1–5	100,000	500,000	100,000 writes x 5 days
Day 6–15	250,000	2,500,000	250,000 writes x 10 days
Day 16–31	45,000	720,000	45,000 writes x 16 days
Monthly total		3,720,000
Monthly total write units		1.39	3,720,000/2,678,400
Monthly billing		$4.35	3,720,000/2,678,400 x $3.135

Reads
Workload period	Daily read workload	Total reads consumed for each period	Details
Day 1–5	100,000	500,000	100,000 reads x 5 days
Day 6–15	250,000	2,500,000	250,000 reads x 10 days
Day 16–31	45,000	720,000	45,000 reads x 16 days
Monthly total		3,720,000
Monthly total read units		1.39	3,720,000/2,678,400
Monthly billing		$0.22	3,720,000/2,678,400 x $0.16

Storage
	Total storage GB read units	Monthly billing	Details
Day 1–31	5	$0.33	5 GB x $0.066

Total monthly billing $4.91 (Write, read, and storage)

Note: The total monthly billing number is rounded. There will be some slight differences if you simply total the numbers in the examples using a calculator.

Different dynamic workloads with the same consumption and costs

On-demand capacity model bills are based on the total number of read and write units consumed by your application over a month. The example below assumes that you have the following two vastly different workloads: one is continuous with various peaks and troughs, and the other is spiky with abrupt pauses. Both consume the same number of write and read units over a month. The areas under each workload curve represent the total number of write and read units consumed and are the same. Therefore, the monthly costs will also be the same.

Example 1: Various continuous write/read operations in a month

Days 1 to 5 have continuous, medium write/read workloads compared to Days 6 to 15 with higher workloads. Days 16 to 31 have continuous, lowest write/read workloads among the three different periods.

Example 2: Unpredictable sudden spike in write/read operations in a month

Days 1 to 5 have one spike with medium write/read workloads. Then, there is a rest period before Days 6 to 15 which have three spikes with higher write/read workloads and shorter rests in between. Days 16 to 31 have the lowest write/read operations with three smaller spikes and longer rests in between.

Metric description

Per the service description (PDF), the metrics are defined as below:

• Write Unit Per Month: Defined as the throughput of up to one kilobyte (KB) of data per second for a write operation over one month, or approximately 2.6 million writes. Each month is deemed to have 744 hours.
• Read Unit Per Month: Defined as the throughput of up to 1 KB of data per second for an eventually consistent read operation (i.e., where the data returned may not be the most recently written data to the database) over one month, or approximately 2.6 million reads. Each month is deemed to have 744 hours. To achieve the throughput of up to 1 KB of data per second for an absolute consistent read operation (i.e., where the data returned is the most recently written data to the database), the equivalent of two Read Units Per Month need to be provisioned.
• Gigabyte (GB) Storage Capacity Per Month: Defined as 1 GB of computer storage space used by a storage filer of the Oracle Cloud Service during a month of the service.

One Write or Read Unit Per Month provides the opportunity to perform 2,678,400 1 KB writes or reads over the course of a month at an interval of one per second. If your table requires multiple writes or reads per second, allocating greater than 1 Write or Read Unit Per Month is required.

Storage capacity is metered and billed based on the number of GB capacity provisioned throughout a month. Capacity can vary within a month; Oracle NoSQL Database will be billed accordingly.

Billing examples

Below are a few billing examples with different workload patterns and provisioned capacities. We will learn how the monthly billing may be calculated using the quickest and easiest methods to understand. We will explore different billing calculations where capacities fluctuate on a monthly or even an hourly basis. Since write, read, and storage capacities provisioned are metered at per-second granularity, users have the flexibility to calculate using other unit prices, e.g., per minute, per hour.

Use monthly and hourly unit prices for the examples. To derive the hourly unit price, divide the monthly unit price by 744 hours. Hourly price is ideal when the workload changes on an hourly basis.

Pricing and metrics for billing examples

Product	Unit Price Per Month	Unit Price Per Hour
Oracle NoSQL Database Cloud—Write	$0.1254	$0.0001685
Oracle NoSQL Database Cloud—Read	$0.0064	$0.0000086

Assumptions

Assume the following for simplicity:

• The record size for write or read operation is 1 KB or less.
• Reads are eventually consistent which are billed at 1 Read Unit per 1 KB operation.
• Writes are billed at 1 Write Unit per 1 KB operation.
• No secondary indexing. Records are accessed through primary key(s).
• A month has 31 days and 744 hours.
• Storage provisioned on Day One stays constant at 25 GB throughout the month.

Monthly billing examples

Below are three examples with different workloads, provisioned capacities, and billing calculation methods.

Example 1: Predictable, continuously consistent workload and provisioned capacities in a month

Assume your application performs a pretty consistent number of write and read operations for the entire month. The workloads are continuous and only slightly vary for the entire month, enabling monthly consistent capacities provisioned for writes and reads. Capacities are provisioned for the workload's small peak at the middle of the month. Assume the storage provisioned on Day One stays constant at 25 GB (with time to live enabled) throughout the month.

Unit Price Per Month is the best method for calculating the monthly billing since the provisioned write and read capacities and storage are consistent throughout the month.

Days 1 to 31 have consistent workload with minimal changes across the month. Provisioned write and read capacities are at 200 KB/sec (200 Write/Read Units) to meet the peak at the middle of the month.

Billing calculation method

Monthly bill = (Write/Read Unit provisioned) x (Write/Read Monthly Unit Price)

Sample billing

Writes
Workload period	Write units	Monthly billing	Details
Day 1–31 (Write 200 KB/sec)	200	$25.08	200 writes x $0.1254

Reads
Workload period	Read units	Monthly billing	Details
Day 1–31 (Read 200 KB/sec)	200	$1.28	200 reads x $0.0064

Storage
	Total storage GB	Monthly billing	Details
Day 1–31	25	$1.65	25 GB x $0.066

Total monthly billing $28.01 (Write, read, and storage)

Note: The total monthly billing number is rounded.  There will be some slight differences if you simply total the numbers in the examples using a calculator.

Example 2: Dynamic, continuous workloads and provisioned capacities in a month

Assume your application performs various write and read operations for different periods within the month. Each workload is different for the following periods:

• Days 1–5: Provisioned 100 Write Units, 100 Read Units for the medium workload at 100 KB/sec.
• Days 6–15: Provisioned 200 Write Units, 200 Read Units for the larger workload at 200 KB/sec.
• Days 16–31 Provisioned 70 Write Units, 70 Read Units for the smaller workload 70 KB/sec.

Assume the write and read capacities changes occur at midnight. Also, assume the storage provisioned on Day One stays constant at 25 GB (with time to live enabled) throughout the month.

Unit Price Per Hour is the ideal method for calculating the monthly billing by aggregating the total hourly write and read capacities provisioned.

Days 1–5 have 120 hours of 100 Write Units, 100 Read Units provisioned. Days 6–15 have 240 hours of 200 Write Units, 200 Read Units provisioned. Days 16–31 have 384 hours of 70 Write Units, 70 Read Units provisioned.

Billing calculation method

Monthly bill = SUM ((Hourly Write/Read Unit provisioned) x (Write/Read Hourly Unit price))

Sample billing

Writes
Workload period	Hourly write units	Hourly write cost	Total hourly cost for x days	Details
Day 1–5	100	$0.0169	$2.02	Hourly cost x 120 hours
Day 6–15	200	$0.0337	$8.09	Hourly cost x 240 hours
Day 16–31	70	$0.0118	$4.53	Hourly cost x 384 hours
Monthly billing			$14.64

Reads
Workload period	Hourly read units	Hourly read cost	Total hourly cost for x days	Details
Day 1–5	100	$0.0009	$0.10	Hourly cost x 120 hours
Day 6–15	200	$0.0017	$0.41	Hourly cost x 240 hours
Day 16–31	70	$0.0006	$0.23	Hourly cost x 384 hours
Monthly billing			$0.75

Storage
	Total storage GB	Monthly billing	Details
Day 1–31	25	$1.65	25 GB x $0.066

Total monthly billing $17.04 (Write, read, and storage)

Note: The total monthly billing number is rounded. There will be some slight differences if you simply total the numbers in the examples using a calculator.

Example 3: Dynamic workloads and multiple provisioned capacities in an hour

This is a special case where different write/read capacities are provisioned within an hour. Assume your application performs two extreme workload changes within an hour. Below is how the cost is calculated when there are more than one write/read unit provisioned within an hour.

• Days 1–5 : provisioned 100 Write Units, 100 Read Units for the medium workload.
• Day 6 (12:00 a.m.–1:00 a.m.):
  • 12:01 a.m.–12:15 a.m. (15 minutes): 20 Write Units, 20 Read Units provisioned for a drop in workload.
  • 12:16 a.m.–1:00 a.m. (45 minutes): 60 Write Units, 60 Write Units provisioned for a slight increase in workload.
• Day 6 (1:01 a.m.–11:59 p.m.) (23 hours): 200 Write Units, 200 Read Units are provisioned for the sudden increase in workload.
• Days 7–15: Provisioned 200 Write Units, 200 Read Units for the large workload.
• Days 16–31: Provisioned 70 Write Units, 70 Read Units for the small workload.

Unit Price Per Hour is the ideal method to calculate the monthly billing by aggregating the total hourly write and read capacities provisioned. In this case, the Unit Price Per Hour is also used to calculate the average cost during the hour that has two changes in write and read capacities provisioned.

Days 1–5 have 120 hours of 100 Write Units, 100 Read Units provisioned. Day 6 (12:01 a.m.–12:15 a.m.) has 20 Write Units, 20 Read Units provisioned. Day 6 (12:16 a.m.–1:00 a.m.) has 60 Write Units, 60 Read Units provisioned. Day 6 (1:01 a.m.–11:59 p.m.) have 200 Write Units, 200 Read Units provisioned. Days 7–15 have 216 hours of 200 Write Units, 200 Read Units provisioned. Days 16–31 have 384 hours of 70 Write Units, 70 Read Units provisioned.

Billing calculation method

Monthly bill = Sum ((Hourly Write/Read Unit provisioned) x (Write/Read Hourly Unit price) + (12:01 a.m.–1:00 a.m. Hourly Average Write/Read Unit provisioned) x (Write/Read Hourly Unit price))

Sample billing

Here's a sample of the monthly billing, assuming the storage provisioned on Day One stays constant at 25 GB (with time to live enabled) throughout the month. The average write/read units in Day 6 between 12 a.m.–1 a.m. is as follows:

• 12:01 a.m.–12:15 a.m. = 20 write/read units * 900 sec = 18,000 write/read units
• 12:16 a.m.–1:00 a.m. = 60 write/read units * 2700 sec = 162,000 write/read units
• Day 6 average write/read units for 12 a.m.–1 a.m. = (18,000 + 162,000)/3600 = 50 write/read units

Writes
Workload period	Hourly write units	Hourly write cost	Total hourly cost for x days	Details
Day 1–5	100	$0.0169	$2.02	Hourly cost x 120 hours
Day 6, 12 a.m.–1 a.m. (Average provisioned capacities)	50	$0.0084	$0.01	Hourly cost x 1 hour
Day 6, 12:01 a.m.–12:15 a.m. (Write 20 KB/sec)	20			20 writes x 900 seconds
Day 6, 12:16 a.m.–1:00 a.m. (Write 60 KB/sec)	60			60 writes x 2700 seconds
Day 6, 1:01 a.m.–12:00 a.m.	200	$0.0337	$0.78	Hourly cost x 23 hours
Day 7–15	200	$0.0337	$7.28	Hourly cost x 216 hours
Day 16–31	70	$0.0118	$4.53	Hourly cost x 384 hours
Monthly billing			$14.62

Reads
Workload period	Hourly read units	Hourly read cost	Total hourly cost for x days	Details
Day 1–5	100	$0.0009	$0.10	Hourly cost x 120 hours
Day 6, 12 a.m.–1 a.m. (Average provisioned capacities)	50	$0.0004	$0.0004	Hourly cost x 1 hour
Day 6, 12:01 a.m.–12:15 a.m. (Read 20 KB/sec)	20			20 reads x 900 seconds
Day 6, 12:16 a.m.–1:00 a.m. (Read 60 KB/sec)	60			60 reads x 2700 seconds
Day 6, 1:01 a.m.–12:00 a.m.	200	$0.0017	$0.04	Hourly cost x 23 hours
Day 7–15	200	$0.0017	$0.37	Hourly cost x 216 hours
Day 16–31	70	$0.0006	$0.23	Hourly cost x 384 hours
Monthly billing			$0.75

Storage
	Total storage GB	Monthly billing	Details
Day 1–31	25	$1.65	25 GB x $0.066

Total monthly billing $17.01 (Write, read, and storage)

Note: The total monthly billing number is rounded. There will be some slight differences if you simply total the numbers in the examples using a calculator.

Sample billing statement

Oracle Cloud Infrastructure (OCI) monthly billing sample for Oracle NoSQL Database Cloud

Here is a sample OCI monthly billing statement for December 2022 based on Example 2 above.

Open image in new window