“Ant Financial Services11.11: Alipay and Ant Huabei’s Technical Architecture and Practice” Thoughts after Reading
I introduced the overall structure of Alipay by reading “Ant Financial 11.11: The Technical Architecture and Practice of Alipay and Ant Huabei”, and took “Ant Huabei” as Example, introduced how the software architecture works, which gave me some understanding of how the software architecture works.
Article link: https://mp.weixin.qq .com/s?__biz=MzAwNjQwNzU2NQ==&mid=401735445&idx=1&sn=0f8cc086e413332fddad5b05f0363a7e&scene=21#wechat_redirect
This article is mainly based on the structure of Alipay, Architectural features, distributed data structure, data reliability and Ant Huabei are described.
The structure of Alipay is mainly divided into three layers:
(1) operation and maintenance platform (IAAS)
(2) Technology platform (< span lang="EN-US">PAAS)
(3) Business platform (SAAS)
Alipay’s architecture is a logical data center architecture. Alipay will face tremendous pressure during the Double Eleven promotion, so the system scale is very large and the complexity is also very high. Therefore, Alipay has a comprehensive and scalable solution that can be expanded in the dimensions of a unit and is no longer dependent. In the same city IDC can realize the N+1 remote disaster recovery strategy, greatly reducing the cost of disaster recovery, and business-level traffic ingress and egress are formed With unified controllable and routable control points, the controllability of the overall system has been greatly improved. The data of the trading system in “Ant Financial 11.11: Alipay and Ant Huabei Technical Architecture and Practice” is mainly divided into three large database clusters:
1. The main transaction database cluster. Each transaction creation and status modification is first completed here, resulting in The changes are then replicated to the other two database clusters through the reliable data replication center: the consumption record database cluster and the merchant query database cluster. The data of the database cluster is split horizontally into multiple copies. In order to ensure scalability and high reliability at the same time, each node will have a corresponding spare node and failover node. When a failure occurs, you can switch to the failover node within seconds.
2. The consumption record database cluster provides consumers with a better user experience and Demand;
3. Merchants query the database cluster to provide better user experience and needs;< /p>
Because Alipay is more sensitive to transaction costs than traditional financial companies, the development of Alipay’s data architecture is a low-cost, linearly scalable, The evolution of distributed data architecture. Now Alipay’s data architecture has been upgraded from a centralized minicomputer and high-end storage to a distributed PC service solution. The overall data architecture solution is as far as possible to be independent and standardized.
“Ant Financial 11.11: Alipay and Ant Huabei’s Technical Framework and Practice” mentioned in the article To “Looking at the current sharing of various architectures, everyone likes to talk about “planning” and sharing of various architecture design scheme optimization strategies, but in reality, there are two situations in the end: “Blowing the cowX has not been verified at all.” (The ability of various frameworks has not been tested at all, just empty talk), “Blowed The cattleX broke through the actual test” (The design concept is very good, but as soon as it encounters the impact of the actual big business, the system will Hung up), there are very few successes in the end. These explanations although the structure of “Spirit Chicken Soup” and “Successful Learning” technicians are already familiar with it, but found that it’s not the case when they practice. The theoretical level of analysis and design in terms of “seeking“, the most important thing is to implement “devices” and “The level of “.”< /span>
In the article, “scheme”, “device”, and “will” refer to the overall architecture design plan and strategy, various basic middleware and basic components that support technical work, respectively And technicians who have grown up through practice. Most people think that architecture design schemes and strategies are more important, but through most successful cases, it is found that there are various infrastructure tools with excellent and high stability and technologies that have been “tortured thousands of times” through battles. The support of personnel is the key to the final victory, and these two aspects cannot be achieved through learning other structures and successful cases, but must be summarized through their own accumulation and the characteristics of the business itself.
In the past, we used an open source or commercial component to share technology to obtain the ability to quickly solve and develop technology, but with the gradual change of business complexity, professionalism, and scale The disadvantages of this method are also obvious:
1. Many components simply cannot meet various technical indicators in large concurrency scenarios;
2. As the business becomes more complex and professional, there is no open source component that can be used directly;
3, “person” itself Experience and ability cannot be passed on.
So now, the way we share technology and business capabilities through the “cloud” is also developing faster and faster.