The basic process of QOS is: classification-strategy-identification-queue-scheduling these steps, let's briefly describe.
The first step of QOS must be to classify data, and data with the same transmission quality must be of the same type. The best-effort method, according to the default rules for data classification, unified service model, integrated service model: during the data transmission process, the same service model is used for transmission on the intermediate nodes. Differentiated service model: between nodes There is no need for signaling interaction. The node processes the data separately. The strategy has nothing to do with upstream and downstream, and only depends on the local.
QOS classification: Classifying is classification. The process is to determine the classification of these messages into each data stream represented by the CoS value according to the trust strategy or according to the analysis of the content of each message. Therefore, the core task of the classification action is to determine Enter the CoS value of the message.
Classification occurs when the port receives incoming messages. When a port is associated with a Policy-map that represents a QoS policy, the classification takes effect on that port, and it affects all incoming messages from the port.
Identifying and prioritizing data packets according to the protocol can reduce latency. Applications can be identified by their EtherType.
(2) TCP and UDP port numbers
Many applications use some TCP or UDP ports for communication. For example, HTTP uses TCP port 80. By checking the port number of the IP data packet, the intelligent network can determine which type of application the data packet is generated by. This method is also called the fourth layer switching, because both TCP and UDP are located in the fourth layer of the OSI model.
(3) Source IP address
Many applications are identified by their source IP address. Because the server is sometimes configured specifically for a single application, such as an email server, analyzing the source IP address of the data packet can identify the application that generated the data packet.
(4) Physical port number
Similar to the source IP address, the physical port number can indicate which server is sending data. This method depends on the mapping relationship between the physical port of the switch and the application server.