Novel Reading Network: How does Tingyun Server ensure the performance of a high-concurrency website with tens of millions of users?

As of December 31, 2014, Xiaoshuo.com has been online for 10 years, with a total of 50 million registered users and more than 300,000 resident authors, covering all provinces and cities in China. As the earliest online reading website in China, Xiaoshuo.com is highly praised by literary novel lovers and domestic online authors for its rich content and good user experience.

Novel reading network's pursuit of performance

Novel Reading Network is an online reading website that uses click volume, VIP reading volume, and user retention rate as key business indicators. Website performance is crucial. The situation of Novel Reading Network is as follows:

1. The total number of registered users is 50 million, the number of resident authors is 300,000, the number of online users is over one million, and the average daily visits are over ten million;

2. It uses the latest LAMP architecture, namely Linux-Apache-MySQL-PHP, which has the advantages of rich Web resources, lightweight, rapid development, universality, and cross-platform;

3. Novel Reading Network has a complete distributed system and separate virtual hosts to support its massive daily visits;

4. Have a professional and efficient operation and maintenance technical team.

In addition, the novel reading website has introduced a third-party application performance management solution - Tingyun Server.

Novel Reading Network: Tingyun Server can effectively improve the efficiency of locating server-side performance problems

The high scalability and extensibility of the distributed system plays a role in resource isolation for the website. When a channel has a large number of visits, the distributed system can prevent other channels from being disturbed, avoiding performance problems. If they occur, quickly finding the cause of the problem is the top priority for every high-concurrency website. Novel Reading Network uses Tingyun Server to implement performance monitoring on its servers. Through Tingyun Server, relevant performance data can be obtained instantly for intelligent analysis, helping operation and maintenance personnel to quickly find network bottlenecks, further improving operation and maintenance efficiency, and achieving the best user performance experience.

Generally speaking, users tend to focus on the processing capabilities of applications, that is, whether they respond normally and whether they are available, but often ignore the performance issues on the server side. When problems occur, users may attribute more reasons to network-level problems, or know that there is a problem with the code but do not know which line of code is causing it or what is causing it; or they may judge the problem as a program exception caused by an architectural defect, but these are just inferences without specific data to support them.

Tingyun Server uses probes implanted in application servers to monitor server performance issues, and monitors application code, relational databases, NoSQL, third-party interface call services, and the server itself, identifying the causes of slow response, unavailability, or errors such as network, code, and database. In the past, users could only speculate on the server-side causes.

Novel Reading Network: Tingyun Server can effectively reduce resource loss

From the perspective of operation and maintenance personnel, if they monitor the server, they often face great difficulties and challenges: for example, they need to spend a lot of energy and money, they need manpower to maintain the server, they need a separate server to carry it, and they need a lot of time to analyze the performance data. At the same time, what makes operation and maintenance and R&D personnel most headache is that even if all the above problems are solved smoothly, the actual implementation will still be limited by human factors, and there will be a large communication cost for operation and maintenance and R&D personnel to cooperate. If performance problems cannot be solved in a timely and effective manner, then they will accumulate and become a big problem.

Tingyun Server provides server-side performance monitoring in a simple deployment and minimalist operation manner. Users only need to deploy probes to understand the reasons why the application slows down or reports errors.

The picture shows that Tingyun Server shortens the performance problem solving process

Novel Reading Network

Founded in 2004, it is highly praised by literary fiction lovers for its unique style and rich content. Relying on the spontaneous recommendations of its members, it currently has nearly 60 million daily visits, 2 million online users per day, 200,000 original works, and ranks 171st in global traffic history.

Listen to the Clouds

As a leading end-to-end application performance management solution in China, Tingyun has four important product lines: Tingyun App, Tingyun Network, Tingyun Server, and Tingyun CDN. From mobile clients to server-side and then to the network level, it helps customers monitor and locate performance issues such as crashes, freezes, connection timeouts, inefficient code, slow interactions, and third-party interface calls in real time. Currently, Tingyun's service customers include more than 500 large companies such as Google, IBM, and BAT. It is also the only company in China that Gartner, a leading global technology consulting company, focuses on in the field of APM.