After a new E-Line service is provisioned, the download rate is tested using FTP software. The download rate does not meet the requirement because the latency is too high. Why?
1. Latency refers to the time required for a packet to be transmitted from one end of a network to another end. In an SDH network, the formula for calculating the latency is as follows: t = (5 μs * k + 250 μs * n) * 2 + 1 ms. K indicates the distance between two ends, in km. n indicates the number of NEs that are passed through. 1 ms indicates the latency of the Ethernet encapsulation and decapsulation between the two ends. Generally, the ping test is used to test the latency between the two ends.
2. The FTP software uses the TCP protocol. The TCP protocol has a sliding window. That is, the transmit end must send the bytes required by the sliding window size and can only send more bytes after receiving the confirmation from the receive end. Example: Assume that the PC at both ends of the E-Line is interconnected, the PC is MTU = 1500 Bytes, and the maximum window number is w = 16. In this case, the transmit end can transmit a maximum of 1500 * 16 = 24000 Bytes = 192000 bits at a time. If the latency is 30 ms, that is, the duration from the time when the data is sent to the time when the data is received at the receive end is 30 ms, the download rate is equal to the maximum number of bits that can be transmitted/latency, that is, the download rate is 192000/30 = 6.4 Mbit/s = 0.8MB/s. Therefore, when the FTP single-thread download is used and the sliding window is fixed, the higher the latency, the smaller the download rate.
3. When testing the download rate, you are advised to use multi-thread download or increase the TCP sliding window.