Large portal website development experience

Because permission and range control are involved, every time a user accesses a page, it is the process of running the program on the background, especially when at least two or more SQL statements are executed.
I know that portal websites cannot do this. They all use static pages to support a large volume of traffic (if it is also a dynamic inventory, it will only wait !). What are the ideas for static page generation and management:

I thought about it and summarized it as follows. I hope you can give me some ideas;
1. You can use JavaScript (or include file) to classify the common parts of the web page to facilitate updates, such as topic menus and navigation. In this way, after updating a column, you only need to regenerate the JS file to update the menu column of the entire website.
2. The index page, including the index page of the subtopic, is generated regularly. According to a rule, the latest article is generated for users to access.
3. The document storage directory is opened on the current day, facilitating management
4. Save the title, keyword, url, and topic attributes of an article to the database, which is mainly used for management, modification, and search.
5. To achieve full-text search, you can consider third-party stuff. In this way, a few articles similar to this Article are included after each article.

Recently, we are working on the design of a large comprehensive portal website background management system (Content Management System ). And planning, a deep understanding of website development technology. From the emergence of the Internet to the present, WebSite development technology has four processes: 1. Static WebSite. The page is static and no database is supported. 2. dynamic WebSite, which is currently the most widely used technology for small and medium-sized websites, that is, database-driven WebSite 3. static dynamic content publishing technology. This is a relatively advanced Web development technology. As user traffic increases With tens of thousands or even more visits on the day, the load on the Web server increases rapidly to reach full load or even overload, greatly reducing The system efficiency even causes the system to be difficult to run, although the upgrade from DBMS and hardware facilities can be properly solved This efficiency bottle strength problem, but this greatly increases the cost. In addition, we need to improve the application in search engines. Search When the engine's "Robot program" encounters a dynamic web page during the search process, such as: News. asp? Year = 2003 & When Month = 12 & Date = 19, the robot encountered "?" And "&", it will enter an endless loop, so the robot program should avoid entering an endless loop and use the "Robot trap program". At this time, it can only search News. asp content on this page. After we adopt the static release technology of dynamic content, News. aspYear = 2003 & Month = 12 & Date = 19 this URL is converted to static News/2003121911062033.htm. the content of this static page is News. the content generated by aspYear = 2003 & Month = 12 & Date = 19 not only increases the probability of being searched, In addition, the system is much more efficient (because when a Web user requests to access this page, there is no need to request the Database System Data). This technology is widely used in large and medium-sized comprehensive websites, such as Sina and Sohu. 4. Cache-based Release Technology (reverse proxy ). Some methods have been used in the past: after the first application access, the generated content is saved as a cache file. When the next request is made, the cached file is read from the cache directory. When the content is updated, the application deletes the content from the cache directory to reduce access to the database. Although this can also carry a relatively large load, it is difficult to separate such a content management and cache system. For another idea, you can find that there is a lot of room for choice between content management and cache mechanisms. It is even possible to take advantage of the comprehensive content management system that is already a "function", and make all "efficiency" problems more professional at the front-end, it is also easy to solve the distribution of cache servers: It can be through open source code SQUID for reverse proxy WEB acceleration, it can be a specialized cache hardware device, or even a professional cache service provider. Dynamic cache must have a Parameter Parsing Process Based on the static link itself. Many professional content management systems use the Parameter Parsing mechanism as a WEB server module. Can we set the previous HTTP/GET method? The key = value is directly transmitted in the/value1/value2 mode, thus realizing the static URL form of the dynamic page. Instead, you only need to add a CACHE server on the front end, such as Squid. Dynamic cache publishing of dynamic website content can be achieved. This is the most advanced website development technology at present. It implements Software Engineering in the true sense and the development process. The real Division of Labor in Development greatly reduces the dependency in the development process and increases the development speed. The artist only processes the Template Design and page planners only focus on Layout Planning. Web programmers only focus on functional module development, while System Integration programmers Only by integrating the call and logic of the page template and program module, the interface and content are completely separated, that is, what you see is what you get. In website management, the artist can maintain the template without understanding the program, the technical Administrator does not consider the interface to process the program.