Chat How to grasp the overall capacity planning of E-commerce website

Before, shared the outline of E-commerce site Capacity Planning, overall, capacity planning is a very complex work, involving various technical areas of professional content. Here today to share some ways to grasp the overall e-commerce site capacity planning.

(1) Divide and conquer

Large e-business systems can be very complex, including data centers in major locations around the world, global wide area networks, local exchange-area networks in data centers, security infrastructure, computing infrastructure for data centers, storage subsystems, distributed database systems, and so on. From the point of view of system design, the design of complex system is usually based on the Top-down modular and hierarchical architecture. Hierarchy is a division based on structure, modular based on a functional division. Complex systems can be based on hierarchical division, and then subdivide functional modules at each level, or you can partition subsystems based on modules, and then divide them within subsystems.

This design method simplifies the complexity of the problem, and better controls the number of factors that need to be considered at each stage of the redesign. E-commerce system capacity planning, corresponding to the decomposition of the system structure, can also divide and conquer, the various levels, the performance of each module capacity planning. Specific to all levels, modules. They have different requirements for expected capacity, performance, and scalability, and divide and conquer can also ensure that capacity planning is considered sufficiently detailed.

In addition, divide and conquer can help to grasp the focus of capacity planning. According to the 8/2 principle, the overall performance of the system bottleneck, often depends on the formation of system 20%, divide and conquer to help find the system needs to pay attention to the level and module, in capacity planning for these levels and modules to focus on, often can achieve a multiplier effect.

(2) using tools

Capacity planning is both science and art, because no one can accurately predict the actual capacity requirements. The tool can help to predict and analyze the behavior and demand development trend of e-commerce system, provide sample data for capacity planning, and provide practical data reference for theoretical calculation.

There are two types of tools that can help with capacity planning. One is the test tool, which is typically used before the system is deployed online and upgraded to simulate the actual actual access requests of user users, and can be used to test and observe the behavior and performance of the system as it approaches real-world access requests. Simulating a certain number of concurrent accesses can be used to perform stress tests to help test the changes, trends, and saturation point of the system's behavior and handling capabilities under different loads. These data can greatly assist in capacity analysis and prediction. Common tools such as Microsoft's free tools Web capacity analysis tool and Web Creator Stress tool.

The other is a detection tool, usually used in the system operation and maintenance phase, used to monitor the actual access and utilization of the system, observe the access requirements and peak and valley value characteristics and trends, to help warn system capacity close to the upper limit need to expand, to help predict the amount of traffic and growth in a certain period Facilitates the planning of further upgrades to the foreseeable future system. Some tools can also automate some predictive analysis, predicting capacity requirements for the coming months, years, and year. 3) Dynamic Adjustment

Even when the system is in a very detailed capacity planning before the system is actually started, it may still be found to be different from the actual situation. In addition, with the rapid development of E-commerce business, system access needs change may soon exceed the expected value. Therefore, capacity planning is a long-term task, charm personnel need to regularly assess the system's load, utilization, access needs changes and trends, if necessary, should be proposed to expand the needs of the upgrade. Regular evaluations can be conducted using data from monitoring tools, or through theoretical analysis, according to new requirements.

Here are some of the key factors to consider in capacity planning.

1 Number of users

Determine the number of users who use the site at the same time, which typically has two main sources: Market analysis and system analysis. If the site has not yet been established or published, site owners and operators may need to use market analysis reports to predict the initial and future traffic expected from the site. If the site has been established and is running, analyze the log files of the Web server to see how many clicks the site has taken at different times, and all usage trends that indicate whether the site's content is more or less popular. The number of users currently supported by the site is calculated based on peak usage rather than on average or average usage based on typical usage.

A usability problem is implied here. Enterprises should make decisions about the degree of usability of the website. For example, is it necessary to ensure that the site can handle peak traffic, or is it enough to handle 80% of peak traffic? It depends on how much money the company intends to invest to ensure that the e-business system is always able to meet demand. It depends on whether the loss of E-commerce Web site is worth the cost of the enterprise to achieve this possibility.

2 page Type

If the Web service has processed static pages primarily, because static page reading is mainly to see disk operations, so more consideration is I/O performance requirements, if the Web services need to deal with a large number of dynamic pages, then the middleware System computing capacity and database processing capacity requirements will be relatively high, The greater the amount of data query processing required to generate dynamic content, the higher the system computing capacity and database processing capacity are required.

3 What kind of Web application technology is used

Using different Web application technologies, such as ASP, ASP.net java, applet, JSP, Web service, etc., the cost of processing capacity of the system is not the same.

4 The content of the service provided by the website

For example, through E-commerce to provide legal advisory services of the Web site, the content of the Web page may be mostly text-oriented, with a small number of pictures, then the same conditions, the user calls a certain page time will be relatively short, each user requests to occupy a system of processing capacity resources will be relatively few. And if it is through E-commerce to provide video, image information Services Web site, the page in the picture, video data mainly, may have real-time transmission of streaming media content, then the user calls a page for a long time, each user request possession system processing capacity resources will be more.

For the basic design of e-commerce sites, each set up, including web and middleware systems, database systems, server systems, storage systems, network system capacity planning, simplified, are based on the formula: System capacity = Concurrent Service requests x per request system overhead.