# Aliyun-ACE-RDS storage
## RDS version
### Basic Edition
- Only MySQL5.7
- single-node deployment
- the cost is very high
- calculating and storing the separation
- for individual learning and testing
- availability is not high
### High Availability Edition
- Classic High Availability architecture (a main one)
- physical servers and local hard drive SSD
- Best performance
- binlog data replication technology
### Financial Edition
- a standby two main
- MySQL5.6
- Multi-level replication over multiple copies binlog log to ensure strong data consistency
- Financial-level data reliability and disaster recovery capabilities across the room.
- High availability and security
## RDS Specifications
###-class property
- Multi-instance deployment
- CPU and storage space multiplexing
- there is competition for resources
### exclusive type
- Multi-instance deployment
- CPU and memory isolation
- will not occur multiple instances of resource competition
### monopoly type
- refers to a MySQL instance occupied by one server,
- the best performance
- the most expensive
## Product advantages
### flexible billing
- pay by volume
- Annual subscription
### ready to use
- no need to purchase server hardware and software
- within a few minutes to create a good console or API
### change with demand
- change the database instance configured according to business needs
### transparent compatible
- RDS is consistent with the use of native database engine method
- RDS compatible with your existing programs and tools
Convenient Management ###
- Ali cloud responsible for the daily maintenance and management of RDS
- Ali cloud console or API complete database of its own increase, delete, reboot, backup, recovery management operations
## storage type
### Local SSD disks
- Local SSD disk means that the database engine is located in the SSD the same node. The data is stored in the SSD locally, can reduce I / O latency
- When the RDS instance belongs is a main one (HA Version) or a two master device (Financial Edition) architecture, the primary node fails, the standby node second stage to complete the handover
### SSD cloud disk
- SSD Cloud resilient disc means block storage device based on distributed storage architecture. The data stored in the SSD cloud disc, i.e., calculation and storage to achieve separation
- distributed cloud disk, ensure data reliability through redundancy multiple copies
### ESSD cloud disk
- Enhanced (Enhanced) SSD cloud disk, Ali launched a new ultra-high performance cloud cloud disk products. ESSD cloud disk-based next-generation distributed block storage architecture, combined with 25GE network and RDMA technology, to provide you with a single disk up to 1 million random read and write capabilities and the ability to lower one-way delay
- distributed cloud disk, ensure data reliability through redundancy multiple copies
## Security Mechanism
### The internet
- IP whitelist, VPC network, SSL (Secure Socket Layer)
### storage
- TDE (Transparent Data Encryption), the automatic backup
### Disaster Recovery
- city disaster recovery (multiple instances available area), Disaster Recovery (two multi-center)
### data encryption
### anti-DDoS ***
### traffic cleaning
### black hole processing
## backup, disaster recovery instances
### backup
- data backup
- periodic backup force, more than two days a week and must select two days backup and backup cycle period to the total amount of regular physical backup.
- backup any time via the console or API to initiate the full amount of temporary physical backup at any time.
- Log Backup
- Optionally, you can choose to turn on or off. If you turn off log backup, you can only restore data to a point in time where the backup set when restoring data
- Data backup and log backup using the same expiration deletion policy. The number of days overdue backup can be set to any number of 7-730, you can also delete older backups in real time by adjusting the expiration policy
### recovery
- Press the backup set recovery
- the data can be restored to a specified backup set temporary instance or instance clone
- by a point in time recovery
- can select a point near time, the system according to the full backup and log backup after the data back onto the instance of a temporary or clone instance.
- coverage restoration
- the specified data can be restored to the current backup set RDS instance, rather than a temporary instance or instances cloning. However, if you use this way of recovery, after the example of the restoration will not have data recovery function, caution. Closely related to data recovery and backup policies
- Early Data Recovery time depends on early data backup (and expiration policies related to the frequency of data backup)
- Data recovery after the night time depends on a log backup (a great relationship with the log generation amount)
- Data Recovery supports recovery according to the schedule depends on the log backup is turned on.
- depending on the speed of data recovery data backup frequency (also very much related to the amount of generated log
### Disaster Recovery
- Multiple instances of available area
- RDS single instance running on the main area available in the same region of the available two physical servers, the available area of the enclosure, air conditioning, circuits, networks have redundancy. Asynchronous / synchronous-half data replication and efficient switching mechanism HA, RDS provides the user database availability limit greater than the physical servers.
- RDS supports multiple instances of the usable region (also called dual-city disaster recovery room or city example) a transparent handover procedure
- Application of handover occurs when the DR disconnects connection to the database, the application needs to reconnect RDS
- Examples of disaster recovery
- to break through the limitations of disaster recovery capability between different regions of the same area available
- Zone A of Example RDS A 'transmission through the asynchronous copy data (Data Transmission) Zone B to the RDS Example B
- Example B 'is a complete stand-alone RDS instances, independent connection address, account number, and permissions
- non-transparent type of switch, change of address application
- DR before handover user needs to stop the instance A 'to an instance of B' data replication, so as to avoid confusion of data
## schema extension
### read-only instance
- scenarios
- Read write little scene, the scene for a conventional separate read and write, read-only instance to meet the demand a large number of database reads, increase throughput applications.
- The principle
- Preparation Example reproduce data from, the data is consistent with the primary instance, after the data updating operation is completed will be the primary example of the master instance automatically and immediately synchronized to all read-only instance, the delay may be provided a read-only copy on read-only instance instance
- the difference between the read-only, example of a backup and disaster recovery
- read only examples and examples standby in the same region, can be used in different areas; Preparation Examples and Examples in the same main region; Examples and disaster recovery standby instances different regions
- read-only instance can be created if necessary up to 10; the number is fixed Preparation Examples 1 to 2, a version of availability, three nodes Enterprise Edition (formerly financial Edition) 2; disaster recovery is a primary example of different regions a prepared
- Create and set read-only instance will continue to provide services after reading weights; standby instance and disaster recovery instances not provide services in the primary instance normal operation
- functional limitations
- the number of instances of read-only
- create up to five read-only instance of MySQL memory <64G
- MySQL memory is large> = create up to 10 read-only instance 64G
- Examples of backup
- Because the master instance has been backed up, the backup does not support read-only instance settings and manual backup
- Examples of recovery
- does not support the creation of backup files or temporary instance through any point of time, it is not supported by examples covering the backup set
- After you create a read-only instance, the primary instance is not supported by the backup set covering instances directly to recover data.
- data migration
- does not support migrating data to a read-only instance
- Database Management
- does not support the creation and deletion of database
- Account Management
- does not support the creation and deletion of the account, the account does not support the authorization and modify account password function
- Features
- the geographical area and the available
- the main examples in the same region, can be used in different areas
- billing
- according to the amount paid, more flexible, cost that is cheaper deductions every hour, read-only instance when the specifications vary depending on deductions
- Specifications
- may be inconsistent with the primary instance, and can be changed at any time. Read-only instance specifications recommended not less than master instance specifications, otherwise easily lead to high latency read-only instance, the phenomenon of higher load
- Network Type
- may be inconsistent with primary example
- Account and Database Management
- no need to maintain an account with the database, all synchronized by master instance
- whitelist
- whitelist will be automatically copied its primary instance of the read-only information when instance is created, but the white list and read-only instance of the main examples are independent of each other. If you need to modify read-only instance of white list, see white list
- Performance Monitoring
- providing nearly 20 Monitor view system performance metrics, such as disk capacity, IOPS, connections, CPU utilization, and network traffic
- Performance Optimization MySQL
- offering a variety of recommendations, such as the storage engine checks, primary keys of the large table check, the index ones, deletions indexing, the user can optimize the application of its own recommendations and to optimize the characteristics of the database