Distributed ID Generating System – Summary

Distributed id generation system–summary

UUID

Simple and easy to use, but when doing database sharding, UUID is not suitable as a partition Film key

For details, see Leaf——Meituan Dianping Distributed ID Generation System

snowflake

The performance is very high. The disadvantage is that it is easy to make mistakes if the time is called back or the time of each instance node is inconsistent.

For details, please refer to Leaf——Meituan review distribution ID generation system

Meituan’s open source Leaf

Support multiple generation strategies of different modes

  1. Number segment mode
    1. This mode needs to build a DB table, and a special service is needed to provide an interface for obtaining an id. There is a network delay
  2. Snowflake mode

In order to pursue higher performance, the Leaf service needs to be deployed through RPC Server, then only the leaf-core package needs to be introduced, and the ID generating API is encapsulated into the specified RPC framework.

See Leaf

sharding-jdbc open source primary key generation component

Simple and easy to use, you can specify workerId or If you don’t specify it, you can import it directly via jar

Baidu’s open-source uid-generator

You need to build a DB table, you need to have a special service to provide an interface to get ID, there is a network Delay

See uid-generator

distributed id for details Generation system-summary

UUID

Simple and easy to use, but when doing database sharding, UUID is not suitable as a sharding key

For details, see< span class="Apple-converted-space"> Leaf——Meituan Dianping Distributed ID Generation System

snowf lake

The performance is very high. The disadvantage is that if the time is called back or the time of each instance node is inconsistent, it is easy to make mistakes.

For details, please refer to Leaf— —Meituan Dianping Distributed ID Generation System

Meituan’s open source Leaf

Support a variety of different model generation strategies

  1. Number segment mode
    1. This mode needs to build a DB table, and a special service is needed to provide an interface for obtaining ID. There is a network delay.
  2. Snowflake mode
  3. li>

In order to pursue higher performance, the Leaf service needs to be deployed through the RPC Server, then only the leaf-core package needs to be introduced, and the API that generates the ID is encapsulated into the specified RPC framework.< /p>

For details, see Leaf

sharding-jdbc open source primary key generation component

Easy to use , You can specify the workerId or not, and you can import it directly through the jar

Baidu’s open source uid-generator

You need to build a DB table, and you need to have a special service to provide ID Interface, there is a network delay

See uid-generator

Leave a Comment

Your email address will not be published.