Database architecture design strategy based on SaaS model (rethink)

The original design of database data Isolation Based on SaaS architecture is nothing more than three design modes: (1) independent database (2) Share Database, independent Schema (3) Share Database. shareSchema, each table has a tenentid (Tenant ID). The advantages and disadvantages of these three models have been mentioned a lot. I am an Alibaba Software Senior in SaaS architecture design.

The original design of database data Isolation Based on SaaS architecture is nothing more than three design modes: (1) independent database (2) Share Database, independent Schema (3) Share Database. share Schema, each table has a tenentid (Tenant ID). The advantages and disadvantages of these three models have been mentioned a lot. the predecessors of SaaS Architecture Design/I/Alibaba Software

The original design of database data Isolation Based on the SaaS architecture is nothing more than three design modes:
(1) independent database
(2) Share Database, independent Schema
(3) Share Database. share Schema, each table has a tenentid (Tenant ID). The advantages and disadvantages of these three modes are mentioned a lot, this is also the story of the latest SaaS Architecture Design book by my predecessors/Alibaba software, including the SaaS Architecture Design Guide launched by Microsoft two years ago. I have been thinking about this issue recently. Is there no other way out?

Among the three models, I believe that the first SaaS architecture will not be adopted, but the competition between the 2nd and 3rd types of architecture is often no one can convince anyone. The two architectures have their own advantages and disadvantages, and both of them are quite distinctive.

2nd types of architecture design, Share database, and independent Schema are suitable for complex MIS applications, with many database associations. The 3rd types of architecture design are suitable for simple website applications, such as online bookstores, obviously, the 3rd Performance architectures are better, but the data isolation permissions are too complex and have high requirements on programmers. The 2nd Performance problems pose high requirements for software system architects. In addition, there are 2nd types of applications that integrate B2B websites and supply chain/invoicing systems. It is almost impossible to comprehensively query the statistical data of all tenants (suppliers/customers, it is difficult for each tenant to share data comprehensively, and each tenant is a data island.

Is there really no better design? Later, in the constant struggle of every day, we finally saw a ray of dawn. Using the features of the database, we may be able to integrate the advantages of the 2nd class and 3rd class designs, this problem can be solved, but the advantages of cross-database are sacrificed. However, the design of the application architecture is not changed much, and the advantages of design 2 and Design 3 are integrated, I personally think that such sacrifice is worthwhile in specific applications.