MaxCompute takes Project as a charged unit. The bill is charged according to three aspects: the usage of storage, computing resource, and data download respectively. You pay for compute and storage resources by the day with no long-term commitments.
The data which is stored in MaxCompute, including tables and resources, is charged according to the capacity size. You pay for storage resources by the day, but MaxCompute samples users storage usage by hour. The specific billing formula is shown as follows:
|1GB to 100GB
|Next 924 GB
|Next 9 TB
|Next 90 TB
|Next 900 TB
|Over 1024 TB
|Free||0.0028||0.0014||0.0013||0.0011||0.0009||please contact us|
If you store 50 TB data in MaxCompute, the bill is charged as follows:
（100GB-1）* 0.0028 USD/GB/day
+（1024-100）GB* 0.0014 USD/GB/day
+（10240-1024）GB* 0.0013 USD/GB/day
+（50*1024-10240）GB * 0.0011 USD/GB/day
As MaxCompute compresses and stores the user data, the charge is based on the capacity size of the data after compressed. So in most of case, the capacity size is different from the data file size that the user counted by himself. The compression ratio is generally about 5 times.
The bill invoicing time is usually within several hours after the current billing cycle ends. The longest time is no more than 6 hours. After the bill is generated, the fee is automatically deducted from your balance.
MaxCompute supports two kinds of charging methods: one is Pay-As-You-Go, and it means each task is measured according to the input size by job cost; the other one is prepayment by CU cost, and it means users can book a part of resource in advance. This way is only supported on Alibaba Cloud BigData Platform.
Currently, MaxCompute supports SQL, UDF, MapReduce, Graph, and Machine Learning jobs.
- SQL task (not including UDF) is charged by MaxCompute.
- MapReduce task will be charged on December 19, 2017.
- The other tasks are not charged.
Note For the charging information about UDF, Graph, and Machine Learning, you can pay close attention on the announcement of Alibaba Cloud.
Prepayment by CU cost is only supported on Alibaba Cloud BigData Platform. You can pay for the entire reserved resources term (one or several months) with one upfront payment. This resource can be called CU (Compute Unit). One CU means 1 core CPU and 4 GB memory.
|1 CU||4 GB||1 CPU||22.0|
Pay-As-You-Go is for SQL tasks and MapReduce tasks.
Pay-As-You-Go for SQL tasks
It means every SQL Task is charged according to
SQLComplexity. After the SQL task is finished, MaxCompute sends its metering information to billing system and charges in the next day. The corresponding charging formula of SQL computing task is shown as follows:
Computing Cost of One SQL = DataInputSize * SQLComplexity * SQL Price
The price is shown as follows:
|SQL Price||0.0438 USD/GB|
DataInputSize: It refers to the data size that a SQL statement actually scans. Most of SQL statements have partition filtering and column pruning, so this value is generally less than the source table data size.
Column pruning: For example, the submitted SQL is “select f1, f2, f3 from t1;”. Only the data size of three columns (f1, f2, and f3) in t1 are charged.
Partition filtering: For example, a SQL statement includes “where ds>”20130101“. The “ds” is a partition column. Then the data size is charged only icluding the read partition, rather than the data of other partitions.
SQL Complexity: First count keywords in SQL statement, and then converts to SQL complexity. The details are shown as follows:
SQL keyword number = join Number + group by number + order by number + distinct number + window function number + max (insert into Number -1, 1)
SQL complexity calculation:
If SQL keyword number is less than or equal to 3, the complexity is 1.
If SQL keyword number is less than or equal to 6, the complexity is 1.5.
If SQL keyword number is less than or equal to 19, the complexity is 2.
If SQL keyword number is greater than or equal to 20, the complexity is 4.
For example, the input SQL statement is shown as follows:
SELECT DISTINCT total1 FROM
(SELECT id1, COUNT(f1) AS total1 FROM in1 GROUP BY id1) tmp1
ORDER BY total1 DESC LIMIT 100;
This SQL includes 4 keywords ( one DISTINCT, one COUNT, one GROUP BY, and one ORDER), so the SQL complexity is 1.5. If the data volume of table “in1” is 1.7 GB, then the actual consumption is shown as follows:
1.7 * 1.5 * 0.0438 = 0.11 USD
The bill invoicing time is usually before 06:00 in the next day. After the computing task successfully ends , the system counts the data size and SQL complexity. After the bill is generated, the fee is automatically deducted from your balance. If the SQL task is unsuccessful, the deduction terminates.
Similarly to storage, SQL computing also charges the data size after compressed.
Pay-As-You-Go for MapReduce
Charging for MapReduce tasks in MaxCompute will be from December 19, 2017. The corresponding charging formula of MapReduce computing task is shown as follows:
Computing Cost of One MR task = TotalTime * MR Price (USD)
The price is shown as follows:
|MR Price||0.0690 USD/Hour/Task|
Calculating Time for one MR task that is executed successfully = Execution time (Hour)*The number of cores that task calls
If one MR task calls 100 cores, and it takes half an hour, the Calculating Time for the MR task is: 0.5 hour*100 cores = 50
After the MR task is finished, MaxCompute sends its metering information to billing system and charges in the next day.
- You are not charged when a task failed to execute.
- The Calculating time does not contain the waiting time for execution.
- You don’t have to pay the extra money for MR task if you have paid for MaxCompute Annually/Monthly package.
You can download data from extranet through MaxCompute Tunnel. The bill is charged according to the downloaded data size. The charging formula is shown as follows:
Download Cost from Extranet/time = Downloaded Data Volume * Download Price
The price detail is shown as follows:
|Download Cost from Extranet||0.1166|
Upload traffic is free.
Download data Volume: It refers to the size of a HTTP body for one download request. The HTTP body which carryies data uses protobuffer encoding, so it is generally smaller than the original data capacity, but greater than the data stored in MaxCompute after compressed.
You are not be charged if you download data through extranet, such as downloading data to the disk of ECS.
To upload data in MaxCompute, you are not be charged, whether from intranet or extranet.
The system judges whether your balance is sufficient to pay the MaxCompute services of 3 billing cycles, based on the average of the latest 24 hours bills for MaxCompute services. If your balance is not sufficient, the system sends email or SMS to remind.
If you enable the balance early warning, you can receive email or SMS once your balance is less than the warnning value. The warning value can be set by yourself.
If the your available balance is less than the bill amount of last billing cycle, then your MaxCompute services are in outstanding payment status. You must recharge within 24 hours, so that your services are not subject to effect. If you are in outstanding payment for more than 24 hours, Alibaba Cloud suspends your MaxCompute services and you must repay all delinquent bills before you continually use your MaxCompute services. The system sends SMS or email to remind you to renew your status within 24 hours.
After MaxCompute services are suspended, all MaxCompute projects are released after retained for 14 days, in which the data is lost and unrecoverable. After you recharge within 14 days, the services are automatically started and continually used. If you are in outstanding payment for more than 14 days, you are regarded to give up all of MaxCompute projects. You are reminded in one day before your projects are released.