At 9 o'clock on January 4, Xi'an Yimatong was paralyzed again. In this regard, the Xi'an Big Data Management Bureau responded that the number of visits to the one-code link is too large, and it is taking measures to limit the flow and gradually open up in the follow-up. During the epidemic period, the health code is an indispensable travel "artifact", a sweep to know the identity information and health status, greatly improving the efficiency and accuracy of the verification work, Xi'an one code pass many collapses have brought great inconvenience to people's production and life.
So who is to blame for the collapse of Xi'an One Code Pass again? The reason for the collapse of Xi'an One Code Pass is not yet clear, but it can be seen from the previous system crash.
At about 7:40 a.m. on December 20, the number of user visits to Xi'an's "One Code Pass" surged, reaching more than 10 times the previous peak per second, resulting in network congestion, resulting in some application systems, including "One Code Pass", being unable to use normally. Liu Jun, former director of the Xi'an Big Data Bureau, put forward the view of "non-essential non-bright code": in the special period of nucleic acid testing for all employees, in order to reduce the pressure on the system, it is recommended that the general public do not need to show the code and bright code, and when the system is stuck, please be patient and try to avoid repeated refreshes.
Obviously, this statement is difficult to convince. In the end, why did Xi'an One Code Pass collapse? In this regard, more than 10 experts from Tencent, Huawei, ZTE, and ICT data analysis analyzed the main reasons from the front-end, back-end, and testing aspects.
1, the flow restriction problem: the citizens in a long time can not brush out the health code, many times quit the refresh retry, new traffic arrived at the server, resulting in server pressure becomes larger, bear the load increased, indicating that Xi'an "one code pass" system did not do a good job of flow restriction measures.
2, server problems: the server is limited by peaks, it is impossible to withstand unlimited concurrency capabilities. The cause of server paralysis is that in the same period of time, the number of visitors is large, resulting in a burst of high traffic, beyond the scope of the server.
3. Architecture problem: Xi'an"one code pass" function affects the "nucleic acid detection" service, indicating that the modules affect each other from interface to data calls, which may not be a microservice architecture.
4, performance overload: typical performance overload scenarios, whether the internal root cause is the database bottleneck point, or the number of network links bottleneck point, etc., external causes are caused by overload.
5. Scenario problem: When the big data query is downloaded, one thread occupies too many resources, resulting in other services waiting and even the nucleic acid information in the personal electronic code is not displayed. Therefore, it is estimated that Xi'an "One Code Pass" is a portal, and data and even "cards" are drawn from various subsystems to the server hanging up and down;
6. Design vulnerabilities: the situation of high traffic and high load is not considered, resulting in insufficient testing; the product design does not consider tens of millions of concurrent accesses, and the same level of stress test is not carried out before delivery.
7. Stress test: In the case that the public cannot see the health code for a long time, the multiple exit refresh retries, and the new traffic arrives at the server, resulting in the pressure on the server becoming larger and the load increasing, indicating that the stress test is not enough.
According to technical experts, the lack of stress testing of the platform by Xi'an Telecom in the early stage was the main reason for the paralysis of this service.
Is the problem of inadequate stress testing really difficult to solve? In contrast, Wuhan's epidemic prevention work is more tested. Wuhan has experienced three major epidemics, and the "speed, stability and orderly" performance of each response to the epidemic is in stark contrast to the epidemic prevention in Xi'an. Take the sudden epidemic in August, in the face of more than 50 million nucleic acid inquiries, Wuhan still successfully won the epidemic prevention and control battle.
So who are the companies involved in the construction of Xi'an One Code Pass, and who should be responsible for this? According to relevant network information, Xi'an Personal Electronic Identification Code (i.e. Xi'an "One Code Pass") was led by Xi'an Big Data Resources Management Bureau, developed and deployed by China Telecom Xi'an Branch, and promoted and used throughout Xi'an on February 25, 2020. According to the official website of Xi'an Big Data Resources Administration, Neusoft, Alibaba Cloud, Venustech, Anheng and many other companies are also involved in the security work of Xi'an "One Code Pass".
According to network information, when three mobile telecom unicom bid for Xi'an one-code pass, Xi'an Telecom won the bid, and then subcontracted the business to three companies - Zhejiang Anheng, Merrill Lynch Technology and another company. It is rumored that the staff in charge of the development and operation of Xi'an Yidian at that time fell into the protagonist of the dead loop of "no green code can not go to work, no work can not repair the software" this programmer is the staff of Merrill Lynch Data.
After all, Merrill Lynch Data is only responsible for data, and there are many reasons for system crashes. It is suspected that the vice president of Merrill Lynch responded to the explanation about Xi'an OneCode Pass in the circle of friends: Merrill Lynch Data participates in the coding algorithm, and Telecom and Neusoft are responsible for operation. This system crash has nothing to do with our company.
There is a lot of speculation about the cause of the problem. In the final analysis, technical support must be sufficient.
According to the official website of the Ministry of Industry and Information Technology on January 4, 2021, from December 30 to 31, 2021, Han Xia, chief engineer of the Ministry of Industry and Information Technology, went to the Shaanxi Provincial Communications Administration to conduct research on epidemic prevention and control work. Han Xia came to Xi'an "One Code Pass" work class to understand the emergency disposal measures of the nucleic acid sampling system, and required the system to be optimized and the details to be perfected to ensure that there was no congestion downtime.
In the early morning of January 5, the Organization Department of the Xi'an Municipal Party Committee released information on the appointment and dismissal of cadres, and Liu Jun, director of the Xi'an Big Data Resources Management Bureau, was suspended for inspection due to ineffective performance of his duties. Liu Xin is a member of the party group of the Municipal Big Data Resources Management Bureau and is the candidate for deputy director (presiding over the work).
It is hoped that after finding the cause of the accident and strengthening technical support, Xi'an Yimatong can resume work as soon as possible, build a solid safety defense line for the production and life of the people, and do not drop the chain at a critical moment.
ICT Industry Trends Annual Conference 2022
Communication Circle 2022 annual event
Waiting for you to "in the same frame"
2021 China ICT Industry Dragon and Tiger List
excellent solution
Blockbuster!
Author: Cheng Linlin
Editor-in-Charge/Layout: Fan Fan
Review: Shen Qing
Producer: Liu Qicheng
I knew you were "watching"