A visual comparison of common data storage disaster technologies

At present the information society, the importance of database is self-evident. This article from the ordinary users and not manufacturers (not talk about RPO, RTO, MDT, MTBF, mttr and so on professional terminology) to examine and compare a variety of data storage disaster technology, hope to help customers in the purchase of solutions less be fooled, less detours, avoid unnecessary economic losses and system accidents.
For the majority of users, the most concern is the following two points:
A. Is there two logically consistent data: if a disaster-tolerant solution has two ' logically consistent ' data, then in the event of a failure, the user data is secure and the availability of the system is guaranteed. Has two ' logically consistent ' data, which is a must-have for a qualified disaster-tolerant solution. Note that this is about ' logical data ', not ' physical data ', what is logical data, what is physical data, please read it patiently.
B. There is no load balancing read/write separation: Load Balancing read/write separation, especially OLTP and OLAP separation, is recognized as one of the most effective means to improve database performance.
In the current market, disaster-tolerant technologies related to Oracle and SQL Server generally have the following:

2. RAID I
   
   Figure 1 RAID I principle
   A. One DB instance (bold red ellipse in figure)
   B. A logical data (medium coarse red ellipse in the figure)
   C. Two physical data (thin red ellipse in the picture)
   D. No load balancing read/write separation

4. Dual-Machine Hot standby
   
   Figure 2 Two-machine hot-standby principle
   A. One DB instance (bold red ellipse in figure)
   B. A logical data (medium coarse red ellipse in the figure)
   C. Two physical data (RAID) (Thin red ellipse in the image)
   D. No load balancing read/write separation

6. Double-machine Double cabinet
   
   Fig. 3 principle of double-machine double cabinet
   A. One DB instance (bold red ellipse in figure)
   B. A logical data (medium coarse red ellipse in the figure)
   C. Two physical data (thin red ellipse in the picture)
   D. No load balancing read/write separation

8. Storage Dual Live
   
   Figure 4 Storage Dual-Live principle
   A. One DB instance (bold red ellipse in figure)
   B. A logical data (medium coarse red ellipse in the figure)
   C. Two physical data (thin red ellipse in the picture)
   D. No load balancing read/write separation

10. Oracle RAC
    
    Figure 5 Oracle RAC principle
    A. Two DB instances (bold red ellipse in figure)
    B. A logical data (medium coarse red ellipse in the figure)
    C. Two physical data (RAID) (Thin red ellipse in the image)
    D. Load balanced read and write separation

12. Oracle DG
    
    Figure 6 Oracle DG Principle
    A. Two DB instances (bold red ellipse in figure)
    B. Two logical data (medium coarse red ellipse on the figure)
    C. Two physical data (thin red ellipse in the picture)
    D. manual load Balancing read/write separation, target can be queried

14. SQL Server Mirroring
    
    Figure 7 SQL Server Mirroring principle
    A. Two DB instances (bold red ellipse in figure)
    B. Two logical data (medium coarse red ellipse on the figure)
    C. Two physical data (thin red ellipse in the picture)
    D. No load balancing read/write separation, Target library unreachable

16. SQL Server AlwaysOn
    
    Figure 8 SQL Server AlwaysOn principle
    A. Two DB instances (bold red ellipse in figure)
    B. Two logical data (medium coarse red ellipse on the figure)
    C. Two physical data (thin red ellipse in the picture)
    D. manual load Balancing read/write separation, target can be queried

18. Dbtwin Dual Live Cluster
    
    Fig. 9 Principle of Dbtwin double live cluster
    A. Two DB instances (bold red ellipse in figure)
    B. Two logical data (medium coarse red ellipse on the figure)
    C. Two physical data (thin red ellipse in the picture)
    D. Fully automatic load balancing read and write separation

The comprehensive comparison is as follows:

Table A Comprehensive comparison of various data storage disaster technologies

In summary, when a failure occurs, if a scenario has ' two real-time consistent logical data ', then the scheme is undoubtedly the most ideal, if only one logical data, although the physical data has two, but because the physical data only kept sectors or blocks or even the volume level of ' Physically consistent ', lack of database transaction logic protection, so the final database integrity is still risky.
From the level of user data security, the following is a high-to-low sort:

1. Highest: Two copies of real-time consistent logical data.
2. Middle: Two copies of logical data, but there is a short delay in the data.
3. Third: A logical piece of data, but there are two physical data.
4. Lowest: One logical piece of data, and only one physical data.

A visual comparison of common data storage disaster technologies