Method for generating unique ID in distributed system

Abstract

The invention discloses a method for generating an unique ID in a distributed system, related to the field of distributed systems. The method for generating the unique ID in the distributed system is characterized by performing segmented assignment on the ID and storing the APPID of the host, the IP address of the host, the timestamp, and the time self-increasing sequence respectively. The generation speed is fast caused by the shift stitching of different data modules; and the required business information can be carried, therefore, the acquisition and distribution of the information can be easily performed among the distributed systems. According to the method for generating the unique ID in the distributed system, the communication with any external storage devices is not performed in the process of generating the ID, therefore, the possibility of errors is greatly reduced, and the generation process is accelerated; the generation of the ID is based on the timestamp, therefore, the generated ID is monotonic and ordered; and the APPID, IP, millisecond time and sequence number in milliseconds are stitched, therefore, the extension can be performed in the whole IP range, and the requirement of applying the distributed deployment more is greatly met.

Description

technical field [0001] The invention relates to the field of distributed systems, in particular to a method for generating a unique ID in a distributed system. Background technique [0002] In the distributed system, it is necessary to distinguish and identify the machines and equipment in the system. In order to identify different machines and equipment in these different systems, it is necessary to establish the identity serial number of each machine and equipment. This serial number is what we often say ID, in a system we need the ID of each machine device to be unique. [0003] Existing typical solutions include: 1. Database self-growth sequence or field; method: use database, unique in the entire database, advantages: simple, convenient, acceptable performance, natural sorting of IDs, disadvantages: single point risk, when the amount of data exceeds the database When accessing capabilities, you will encounter scaling and performance bottlenecks. 2. UUID, method: direc...

AI Technical Summary

Problems solved by technology

[0003] Existing typical solutions include: 1. Database self-growth sequence or field; method: use database, unique in the entire database, advantages: simple, convenient, acceptable performance, natural sorting of IDs, disadvantages: single point risk, when the amount of data exceeds the database When accessing capabilities, you will encounter expansion and performance bottlenecks

2. UUID, method: directly call the UUID function or generate a third-party library, advantages: simple and convenient; disadvantages: IDs are not sorted and have no business relevance

3. snowflake algorithm: method: use 41bit as the number of milliseconds, 10bit as the ID of the machine (5 bits are the data center, 5 bits of the machine ID), 12bit as the serial number within milliseconds (meaning that each node is in every millisecond It can generate 4096 IDs), and there is a sign bit at the end, which is always 0; advantages: fast speed, high concurrency, and scalability; disadvantages: since there are only 10bit machine IDs, it cannot meet the requirements of multi-agent deployment, and can only be used as a separate ID generation service, so that the application requesting the ID depends on the external network and cannot have enough business information

[0004] The current unique ID generation scheme generally has the following serious problems: high resource consumption, the scheme itself relies on external storage devices, resulting in a large number of IO

There is no business correlation, the service generates ID uniformly, and the application requesting the ID has the dependence of the external network, and cannot be optimized and expanded according to its own business needs

[0005] When the application scenario is strongly related to the business, such as when it needs to be deployed together with the business application, we need the generated ID to carry the corresponding information as much as possible; do not interact with the external network, the unique ID is directly generated locally, and has low resource consumption. When generating functions such as fast generation speed, high concurrency, and high scalability, the existing methods for generating unique IDs cannot meet the requirements

Images