Everyone is very concerned about a topic. There is a Mozilla Optimizer before, but this optimizer has some options to easily cause the page to display misal, so the administrator cancels the top. So, how to optimize Firefox's browsing speed, display speed? Is Firefox not multi-threaded browsing? (There are many people saying that Firefox is relying on multithreading to improve the browsing speed) to turn a post to explain these questions ------- Before the post ------- First Optimization option: Nglayout.initialPaint .delay, if I don't make a remember, this parameter role is: explain the web page and display the previous delay time, unit milliseconds. Everyone knows that the webpage should be displayed, it must be content, which is page data obtained from the server, and obtaining the web data requires time, so this parameter is used to get the web page process for time. So if you set it to 60000 (one clock), after waiting for a clock, the web page is completely displayed; if set to 100 (default), then some webpage has been downloaded this time, Part is shown, that is, the web page is gradually displayed. Luo Yu for a long time, I don't know if I understand this value. This parameter is not too great to increase the browsing speed. 100 is enough. --------- Posted: original location: http: //forum.mozcn.org/viewtopic.php id = 14 Author:? Boson ------- About firefox inside and HTTP connection-related parameters Setha, first moving the ATPPP, sending a letter: ATPP (BIG MOUSE), the letter area: Firefox Title: Parameter settings related to Firefox and HTTP connections: BBS Shuimu Tsinghua Station (Sun Nov 21 08:14:43 2004) The parameter setting related to Firefox and HTTP connections is this paper. This article is the translation of my chat and a big bull. It is a popular, some places may be not accurate. You don't want to be a professional technical article. Welcome to shoot bricks. A typical web page is composed of an HTML file and embedded elements, including images, CSS files, JavaScript files, and more. Each embedded element is not distinguished from the HTTP protocol and the HTML file: that is, you need a browser to go to the server. An earlier typical browser is implemented: When the user is knocked into the URL, the browser and the server establish a connection, request this HTML page, then receive the HTML page sent by the server, follow the in-inline element, can immediately Open a second connection request. In addition, if the embedded element is much, he may open a plurality of connections. When all the elements you need are downloaded, the browser will draw the page. This process is the browser assigned by the earliest HTTP / 1.0 protocol. HTTP / 1.0 This multi-connection mode of operation is improved.
The process of establishing a TCP connection is this: the client sends a network package to the server says that I have to connect with you. After receiving the server, I will return to a network package to say "I am willing", and the client should send it to the server a network package. "Good, let's start the data." This will go to three packages to establish TCP connections. After the connection is established, the browser sends a request, and the server responds to the browser. After the end, you have to go back and forward several network packets to close the TCP connection. If the page has a lot of file lengths, each element requires a single connection that will cause a large number of TCPs to establish a connection and disconnect network packages on the network. In addition, TCP has a characteristic called Slow Start, and its meaning can be substantially thus explained: TCP connection requires the sender to send a certain number of network packages to the receiving end to return to a "I received" network package, and the network package is over When the router is rewritten, the larger the network package, the higher the network, the higher the network, the higher the network. The method for TCP connection finds the optimal network package size is that at the beginning of the TCP connection, the size of the network package is small. According to the network, the two-end program will gradually increase the size of the network package to adapt to the bandwidth to improve network transmission. s efficiency. So the browser sends a request. If the connection is turned off, then the data transfer of this connection is difficult to achieve the speed of the bandwidth. Based on this reason, HTTP / 1.1 quickly came out, put forward the concept of persistent connection, which means that the same HTTP connection can handle multiple requests at the same time, and use certain mechanisms to ensure that each request is Separate. The specific operation process is: After the server sends a response to the browser, it is not immediately turned off immediately; the browser determines that the response of the previous request has been received, and the second request can be sent on this. This mode of operation has greatly reduced the network package, and the experiment also shows that this practice is effective. However, since the server maintains a certain resource on the server, the general server will not be permanently connected, and there is no recommendation of excessive persistent connections between the browser and the server. The persistent connection can be further accelerated. This is pipelining. As can be seen above, the browser needs to wait for a long-lasting response to send the following request. If the server is relatively slow, it is often spent at most of the time for a long time, not data transmission / reception. PiPelining means that the browser can send multiple requests to the server once in a persistent connection, and the server responds to these requests in this connection. This method of operation is especially effective when the browser cache is combined. For example, the picture will exist in the browser cache. When the request is requested, the browser says the server, I already have this picture here, the modification time is xxxx, if this picture is not modified after this picture, Needless to retransmit it. In this case, the server will send a short 304 Not modified type response. If there is no pipelining, every time you ask, you have to wait for the network to transfer to play a back; if there is a pipelining, the browser can ask the server at the same time, if there is a modification, if the server is good for pipelining, it can even Putting four responses to the same network package back, this is a big acceleration. Pipelining is earlier, there is also an imaginary usage. If the server supports Pipelining support, you can put two requests inside the same Pipeline to process, so that the response speed can be further accelerated. Of course, this may not use it.
Ok, come back to see the Firefox About: config network.http. * NetWork.http.keep-alive default is whether TRUE is allowed to last for persistence, this default is True, change to false is a big fool. Network.http.keep-alive.timeout defaults to the 300 persistent connection allowed to keep time, this zoning is meaningless, because the typical server setting is 300. SERVER will have any way you can do you. NetWork.http.max-Connections-Per-Server defaults 8 Connect the maximum number of connections allowed by the same server, which is generally considered to make this numerical value in the case of opening a lasting connection, and not morality. The situation that needs to be protected is more than 10 large files from the website at the same time. NetWork.http.max-persistent-connections-per-server defaults 2 Connect the maximum lasting connection allowed by the same server, this value HTTP / 1.1 standard is 2. Moted up your own network consumption, and generally a server allowed lasting number of persistent connections is limited, you can cause the reduction of others to use, if everyone is large, it means the loss of network efficiency. I personally recommend not moving this value. NetWork.http.pipelining default is whether false allows PIPELINININININING, this feature is not open by default because it is currently still experimenting. It is highly recommended to open. Network.http.pipelining.maxRequests default is 4 Each persistent connection allows the number of requests to be sent. If there is a large image or a long script in PIPELINE, the request that has been sent will be blocked (note the server must respond to requests in turn); and in this case, if not using pipelining, the browser finds one The request processing time is long, and it will naturally use another persistent connection to use as a follow-up request, even further open non-persistent connections. Also, if the server supports PIPELININININININings, the browser is bound to resend the request. Based on this reason, some people think that this number is set to 2 big points to reduce the browsing speed. My personal recommendation is that this value can be maintained by default 4. If the website has a large number of static small pictures, or the network speed is slow, you can try it. Network.http.max-personistent-connections-per-proxy default is the maximum lasting connection allowed by each proxy server. 4 is currently recognized as the most appropriate value, although the recommended value of HTTP / 1.1 is 2. Network.http.proxy.keep-alive default is whether the TRUE Connection Agent server allows persistent connections. True is very good. Network.http.proxy.pipelining The default is whether the False Connection proxy server allows PIPELINING. At present, most proxy servers support Pipelining, so it is generally not recommended. Pipelining is currently a controversial, still in the experimental phase. Although it may really speed up the browsing speed, this depends to a certain extent on the factors of the network, so do not blindly set the relevant parameters in accordance with the online suggestion.
- ※ Modification: · ATPPP 21 10:30:44 Modified this article · [From: 128.12.181. *] ※ Source: · BBS Shuimu Tsinghua Station Smth.org · [from: 128.12.181. *] - -------------------------------------------------- ------------------------- If you are too lazy to see so long, you can modify the following: Enter About: config in the address bar, then find the following All, if you don't have a new NGLayout.initialPaint.delay 100 network.http.keep-alive true network.http.keep-alive.timeout 300 network.http.max-connections-per-server 8 network.http.max-personistent -connections-per-server 2 network.http.pipelining true network.http.pipelining.maxRequests 4 (preferably not more than 8) Network.http.max-personistent-connections-per-proxy 4 network.http.proxy.keep- Alive True Network.http.proxy.pipelining False (If you are willing to try, you can also use true) last time by Mugedy in 2005-12 Wednesday, 16:20, a total of 1 time
