Shulou(Shulou.com)06/01 Report--

This article mainly shows you "how to develop Java performance tuning standards", the content is simple and easy to understand, organized clearly, I hope to help you solve doubts, let Xiaobian lead you to study and learn "how to develop Java performance tuning standards" this article bar.

Why Performance Tuning?

An online product that hasn't been tested for performance is like a timer. When does tuning begin?

Having solved the question of why performance optimization is needed, a new question arises: if a comprehensive performance monitoring and optimization of the system is required, when do we start to intervene in performance tuning? Isn't it better to intervene sooner?

In fact, in the early stage of project development, we do not need to pay too much attention to performance optimization, which will make us tired of performance optimization, not only will not bring improvement to system performance, but also affect the development progress, and even get the opposite effect, bringing new problems to the system.

We only need to ensure effective coding at the code level, such as reducing disk I/O operations, reducing the use of contention locks, and using efficient algorithms. When it comes to more complex business, we can make full use of design patterns to optimize business code. For example, when designing commodity prices, there are often many discount activities and red envelope activities. We can use decoration mode to design this business.

After the system coding is complete, we can test the performance of the system. At this time, the product manager will generally provide online expected data. We will conduct pressure test on the reference platform provided, and count various performance indicators through performance analysis and statistical tools to see whether they are within the expected range.

After the successful launch of the project, we also need to observe the system performance problems according to the actual online situation, log monitoring and performance statistics logs. Once problems are found, we must analyze the logs and fix them in time.

What reference factors are there that reflect the performance of the system?

Above we talked about how performance tuning is involved in various stages of project development, in which performance indicators are mentioned many times, so what are the performance indicators?

Before we look at the performance metrics, let's first understand which computer resources are the performance bottlenecks of the system.

CPU: Some applications require a lot of computation, they will occupy CPU resources for a long time, uninterrupted, resulting in other resources unable to compete for CPU and slow response, resulting in system performance problems. For example, infinite loops caused by code recursion, backtracking caused by regular expressions, frequent FULL GC of JVM, and a large number of context switches caused by multithreaded programming can lead to busy CPU resources.

Memory: Java programs generally allocate memory through the JVM, mainly using heap memory in the JVM to store objects created by Java. The system heap memory reads and writes very fast, so there are basically no read and write performance bottlenecks. But because memory costs are higher than disk, memory storage space is very limited compared to disk. So when the memory space is full and the object cannot be recycled, it will lead to memory overflow, memory leakage and other problems.

Disk I/O: Compared with memory, disk storage space is much larger, but disk I/O read and write speed is slower than memory, although the introduction of SSD solid state disk has been optimized, but still can not be compared with memory read and write speed. Networks: Networks are also critical to system performance. If you have purchased cloud services, you must have experienced the link of choosing network bandwidth size. If the bandwidth is too low, the network will easily become a performance bottleneck for systems with large data transmission or large concurrency.

Exception: In Java applications, throwing exceptions requires building an exception stack to catch and handle exceptions, which consumes system performance. If exceptions are thrown at high concurrency and exception handling continues, the performance of the system suffers significantly.

Database: Most systems use databases, and database operations often involve disk I/O reads and writes. A large number of database read and write operations will lead to disk I/O performance bottlenecks, which in turn will lead to latency of database operations. For a system with a large number of database read and write operations, database performance optimization is the core of the whole system.

Lock contention: In concurrent programming, we often need multiple threads to share the same resource for read and write operations. In this case, in order to maintain the atomicity of the data (that is, to ensure that the shared resource is not modified by another thread when it is written), we will use locks. The use of locks may introduce context switches, which can impose performance overhead on the system. After JDK 1.6, Java has made several optimizations to internal locks in the JVM in order to reduce context switches caused by lock competition, such as adding biased locks, spin locks, lightweight locks, lock coarsening, lock elimination, etc. How to use lock resources reasonably and optimize lock resources requires you to understand more operating system knowledge, Java multithreaded programming basics, accumulate project experience, and deal with related problems in combination with actual scenarios.

Knowing these basics, we can get the following metrics to measure the performance of general systems.

1. Response time

Response time is one of the important indicators to measure system performance. The shorter the response time, the better the performance. Generally, the response time of an interface is in milliseconds. In the system, we can subdivide the response time from bottom to top into the following:

Database response time: The time consumed by database operations, often the most time-consuming in the entire request chain;

Server response time: the time consumed by the server including requests distributed by Nginx and the time consumed by the server program execution;

Network response time: This is the time consumed by network hardware to parse the transmitted request during network transmission.

Client response time: For ordinary Web and App clients, the consumption time is negligible, but if your client embeds a lot of logic processing, the consumption time may become longer, thus becoming a bottleneck of the system.

2. Throughput

In the test, we often pay more attention to the TPS (transaction per second) of the system interface, because TPS reflects the performance of the interface, the larger the TPS, the better the performance. In a system, throughput can also be divided into two categories from the bottom up: disk throughput and network throughput. Let's start with disk throughput, there are two key metrics for disk performance.

Next, look at network throughput, which refers to the maximum data rate that a device can accept without frame loss during network transmission. Network throughput is not only related to bandwidth, but also closely related to CPU processing power, network cards, firewalls, external interfaces, and I/O. The size of throughput is mainly determined by the processing power of the network card, the internal program algorithm and the bandwidth size.

3. Allocation and utilization rate of computer resources

Resource utilization is usually represented by CPU utilization, memory utilization, disk I/O, and network I/O. These parameters are like a wooden bucket. If any of them is short, any one of them is unreasonable, and the impact on the performance of the whole system is devastating.

4. Load bearing capacity

As the system pressure rises, you can observe whether the rising curve of the system response time is gentle. This indicator gives you intuitive feedback on the limits of the load pressure that the system can withstand. For example, when you stress a system, the response time of the system increases as the number of concurrency increases, until the system reaches its limit when it can't handle so many requests and throws a lot of errors.

That's all for "How to set Java performance tuning standards." Thank you for reading! I believe that everyone has a certain understanding, hope to share the content to help everyone, if you still want to learn more knowledge, welcome to pay attention to the industry information channel!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.