Select the backup schedule type - Data Management - Alibaba Cloud Documentation Center

Data Disaster Recovery offers logical backup and physical backup, and provides various backup schedule types to ensure effective data protection and cost-efficient backup services for databases. You can select the backup method and backup schedule type based on your business requirements.

Differences between logical backup and physical backup

Item	Logical backup	Physical backup
Definition	Logical backup backs up database objects such as tables, indexes, and stored procedures. Commonly used tools include MySQL mysqldump and Oracle exp and imp.	Physical backup backs up database files in operating systems. Commonly used tools include MySQL XtraBackup and Oracle RMAN.
Full backup	Data Disaster Recovery splits the data of all tables in a database and executes SQL statements on the database to read the data in multiple threads in parallel.	A backup gateway is installed on a database server to back up database files. For more information, see Install a backup gateway.
Incremental backup	Data Disaster Recovery reads logs stored in the memory of a database and performs incremental backup based on the logs in real time. This prevents an I/O performance drop that may occur when a full backup is performed. The speed at which database logs are read is adjusted based on the rate at which the logs are generated.
Functionality	Finer-grained backup can be performed on databases and tables. Databases are not locked during data backup. This minimizes the impact of data backup on database performance.	A database instance can be backed up only as a whole. Data is read from the disks of a database, which affects the I/O performance of the database. In most cases, physical backup and restoration are faster than logical backup and restoration. Note You can create sandbox instances to use your backup data with ease. For more information, see DBS sandbox overview.
Recommended database size	1 MB to 1 TB	Larger than 1 TB
Supported database engines	Databases created on Alibaba Cloud support only logical backup. For information about the backup methods supported by different database engines, see Database engines and features.
Installation of a backup gateway	Not required.	Required.
Backup data size	Compared with physical backup, logical backup data occupies less storage because logical backup can back up specified tables or data in a database.	The size of a physical backup data is related to the actual storage of the database because physical backup backs up the entire database. In most cases, the size of physical backup data is large.

Select a backup schedule type

Data Disaster Recovery provides various backup schedule types, including serverless, micro, small, medium, large, and xlarge. The free quota for data backup, unit price, and backup and restoration performance vary based on the backup schedule type. A backup schedule type with higher specifications offers higher performance at a lower unit price.

Backup schedule types

The following table describes the specifications of backup schedule types.

Backup schedule type	Estimated time required to back up 100 GB of data	Free quota for data backup (GB per month)	Configuration fee for the Chinese mainland		Configuration fee for the China (Hong Kong) region and regions outside China
Backup schedule type	Estimated time required to back up 100 GB of data	Free quota for data backup (GB per month)	Configuration fee (USD per month)	Unit price of excess backup data (USD per GB)	Configuration fee (USD per month)	Unit price of excess backup data (USD per GB)
micro	20 hours	40	5	0.117	7	0.175
small	5 hours	400	22	0.055	33	0.083
medium	2.5 hours	800	35	0.039	53	0.059
large	1.5 hours	1600	56	0.024	84	0.036
xlarge	1 hour	∞	140	Free of charge	209	Free of charge

Note

If the databases and tables that you want to back up involve issues such as unreasonable table schemas, large tables, or large fields, and you select a backup schedule type with low specifications, the resources of the backup schedule may be insufficient to back up the databases or tables and backup exceptions may occur.
In this case, we recommend that you select a backup schedule type with a high specification when you purchase the backup schedule, or upgrade the backup schedule after an exception occurs, and then try again. For more information, see Upgrade a backup schedule. If the issue still exists, you can join the DingTalk group (ID: 35585947) for technical support.
The fees displayed on the buy page prevail.
Higher specifications offer higher backup and restoration performance at lower unit prices. For more information, see Performance tests on backup and restoration.
You can upgrade or downgrade a backup schedule. For more information, see Upgrade a backup schedule.
In each calendar month, you are allocated a free quota for data backup. If you do not use up the free quota within a month, the remaining free quota does not roll over into the next month.
If the total amount of data that you back up during a month exceeds the free quota, the excess data is charged based on the corresponding unit price.

Select a backup schedule type

You can select a backup schedule type based on your requirements for backup and restoration performance and cost efficiency.

High performance:
If you need to quickly back up or restore data for a database instance such as a database in the production environment, we recommend that you select the xlarge or large type that provides higher backup and restoration performance.
Cost efficiency:
If you do not require high backup and restoration performance and want to select a backup schedule type that is the most cost-efficient, perform the following steps:
1. Estimate the total amount of your backup data, including full and incremental backup data, in a month.
2. Calculate the fees for different backup schedule types and select the most cost-efficient backup schedule type. The following formula is used to calculate the backup fees for a subscription backup schedule:
  Backup fees = Configuration fees + (Actual backup data size - Free quota for data backup) × Unit price of excess backup data
  Note
  For more information, see the Backup schedule types section of this topic.
  For example, Company A plans to perform a full backup for a 150 GB database four times each month. The estimated total backup data size in the current month is 600 GB (150 GB × 4). Company A has the following choices. The prices are applicable in the Chinese mainland.
  - Fees for the micro type: USD 5 + (600 GB - 40 GB) × USD 0.117 per GB = USD 70.52
  - Fees for the small type: USD 22 + (600 GB - 400 GB) × USD 0.055 per GB = USD 33
  - Fees for the medium type: USD 35. The actual backup data size is less than the free quota. No additional fees are charged.
  In this example, the small type has the lowest price. However, if you require better backup and restoration performance, we recommend that you purchase a type with higher specifications.
  Note
  The preceding example is for reference only. In practice, you must also consider the growth of full backup data, such as a 2% increase in the amount of backup data between two consecutive full backups, and the amount of incremental backup data.
  If the total amount of your backup data in a month is greater than 5,207 GB, the xlarge type is the most cost-efficient option. This is because the xlarge type does not have an upper limit on the amount of backup data and offers the optimal backup performance.

Performance tests on logical backup and physical backup

Logical backup and restoration

Test procedure

Prepare an ApsaraDB RDS for MySQL instance. Create a logical backup schedule in Data Disaster Recovery and start full backup and incremental backup tasks to back up data from the ApsaraDB RDS for MySQL instance to Data Disaster Recovery. Then, restore the backup data to the ApsaraDB RDS for MySQL instance. Check the performance of data backup and restoration during this period.

Test environment

Item	Description
Database type	General-purpose ApsaraDB RDS for MySQL
Memory	2,400 MB
CPU	8 vCPUs
IOPS	1,200

Test data

Test data	Description	Example
Database size	The total size of the test database. Unit: GB.	102 GB
The total number of entries.	The total number of entries in all tables in the test database.	0.15 billion
Single record size	The size of one record in the test database. Unit: KB.	1 to 100 KB
Number of fields	The number of fields or columns in a record in the test database.	3 to 22 columns
Field type	The field types in a record in the test database. Basic MySQL field types such as INT, VARCHAR, and DATETIME are included.	Basic MySQL field types.

Test results

Full backup

Backup schedule type	Record per second (RPS)	MByte per second (MBPS)
large	42,855.7	15.3
medium	33,122.2	11.8
small	9,569.3	3.4
micro	6,756.2	1.9

Note

If the source database contains a large number of tables, a lot of time is taken to initialize the tables during backup and restoration. We recommend that you specify a source database that contains no more than 10,000 tables.
The backup process also takes a long time if a table in the source database does not have a primary key, if the primary key of the table is of the String type, or if the table uses a composite primary key.

Incremental backup

Backup schedule type	MBPS
large	46.1
medium	29.8
small	14.9
micro	5.0

Full restoration

Backup schedule type	RPS
large	34,190.5
medium	19,740.9
small	9,949.4
micro	4,320.2

Incremental restoration

Backup schedule type	RPS
large	35,546.9
medium	21,331.4
small	10,061.5
micro	4,972.1

Physical backup and restoration

Test procedure

Prepare a self-managed MySQL instance. Create a physical backup schedule in Data Disaster Recovery and start the full backup task to back up data from the MySQL instance to Data Disaster Recovery. Then, restore the backup data to a specified directory. Check the performance of data backup and restoration during this period.

Note

Data Disaster Recovery reads physical files of the MySQL instance in streaming mode, and uses multiple streams to concurrently write backup data to the built-in storage. Data Disaster Recovery allows you to adjust the number of concurrent backup streams based on specifications. Higher specifications provide a faster backup speed. Two compression algorithms are supported: gzip and LZ4. gzip provides a higher compression ratio and LZ4 provides a faster backup speed.

Test data

Test data	Description	Example
Database size	The total size of the test database. Unit: GB.	40.3 GB
The total number of entries.	The total number of entries in all tables in the test database.	2 billion
The number of tables.	The number of tables in the test database.	160
Single record size	The size of one record in the test database. Unit: KB.	0.2 KB

Test results

Full backup

Backup schedule type and compression format	Time consumed	Backup speed	Size of compressed file
small (4 threads) gzip	636s	63 MB/s	21.1 GB
large (8 threads) gzip	341s	118 MB/s	21.1 GB
xlarge (16 threads) gzip	204s	197 MB/s	21.1 GB
small (4 threads) LZ4	268s	150 MB/s	31.1 GB
large (8 threads) LZ4	119s	338 MB/s	31.1 GB
xlarge (16 threads) LZ4	104s	387 MB/s	31.1 GB

Full restoration

Backup schedule type and compression format	Amount of compressed data	Time consumed	Restoration speed (for the original data amount)
small (4 threads) gzip	21.1 GB	320s	126 MB/s
large (8 threads) gzip	21.1 GB	161s	250 MB/s
xlarge (16 threads) gzip	21.1 GB	86s	468 MB/s
small (4 threads) LZ4	31.1 GB	408s	99 MB/s
large (8 threads) LZ4	31.1 GB	208s	194 MB/s
xlarge (16 threads) LZ4	31.1 GB	108s	373 MB/s