Shardingsphere database sub database sub table introduction

One 、 brief introduction
shardingsphere It is a set of open source distributed database middleware products . It has two important products sharding jdbc and sharding proxy

1.sharding jdbc ： You can see clearly from the picture , Is embedded in the application , All database and table logic is controlled by developers , Relatively flexible functions , But it is more intrusive to the business . Can be understood as an enhanced version of JDBC drive , Fully compatible with JDBC And all kinds of ORM frame

2.sharding proxy: It can be clearly seen from the figure , He is deployed independently , At present, we support mysql and postgresql. Zero intrusion into business , It's like docking with one mysql service , I can't feel his existence , But that's exactly what happened , The function is not flexible and monotonous .（ similar mycat）

Two 、 The core concept
1. Logic table ： Define logical library name and logical table
2. True table ： The partition is really a physical table
3. Data nodes ： Data source and data table
4. Binding table ： The partition rules must be completely consistent with the master sub table （ Table of primary and foreign keys , Very important , The rules of fragmentation must be consistent ）
5. Broadcast table ： There are tables in each partition （ Redundant tables such as dictionary tables ）
6. Patch key ： Specify a field as fragment creation （ It's usually a primary key ）
7. Sharding algorithm ： Distribute the data to different slices by algorithm , Support =,between and in Fragmentation , Developers customize the implementation , High flexibility
8. Fragmentation strategy ： The real slicing operation is the slicing key + Sharding algorithm .

3、 ... and 、 Advantages and disadvantages
Although the problem of single machine capacity is solved by dividing databases and tables , But it also brings many problems ,
for example jdbc Big business intrusion , Business sql It is no longer simply running on the machine , For a large number of trans latitudes join, polymerization , Subquery . Sorting and other functions are difficult to support . Greatly weaken the function of the database . So generally, if you can't divide it, you can't divide it , If you can use distributed database, you can use distributed database , Or add cache ,es And so on .