Network latency is one of the indicators that can not be ignored in evaluating server performance. Low latency means faster server response. How much latency is normal for US cloud servers and will not cause anomalies in server usage and website access? The following is a brief analysis.

In general, network delay and packet loss are caused by network congestion. In order to obtain more accurate ping values, it is best to test when the cloud server in the United States is idle. In this way, the average value of ping tests is obtained for multiple times at different time points, which can reflect the actual performance of the server.

What is the normal latency level for cloud servers in the US?

The Ping operation is to send an echo request to the target US cloud server, and the US cloud server receives the response. Ideally, this round-trip process takes at least 150ms, and if you add the routing and processing time of the cloud server in the United States, the ping value from China to the United States is about 170ms. Choose the US cloud server delay within 200ms is relatively fast. Sometimes packets may have to go through multiple routes to enter the submarine cable, in which case the delay of more than 200ms is acceptable.

In the United States, the California machine room is the closest to China, and the minimum delay is usually controlled at about 170ms. Testing the network latency of US cloud servers can be done using ping, traceroute, various professional network testing tools, third-party monitoring services, multi-location testing, multi-point in time testing, using iprf, etc.

In Linux, open the terminal/command prompt, run the ping command, and add the IP address or domain name of the cloud server. Observe the return response time, which is the network latency from the server to the US cloud server. If the traceroute command is used, enter the traceroute command and add the IP address or domain name of the cloud server to display the path from the server to the cloud server and the delay of each jump point.

There are a variety of online tools and software to help test network latency, such as the MTR tool that combines traceroute and ping.

More detailed analysis and reports can be read with third-party services such as Jtti's free network monitoring and server testing services.

If the service is oriented to globalization, the delay of the cloud server needs to be tested in multiple regions, and more comprehensive delay data can be obtained by testing at different time periods. In addition to the network delay, you can also check the bandwidth usage and packet loss rate. A high packet loss rate may affect service stability.

You can use iprf or speedtest tools, which can test network bandwidth and latency. For example, iperf can measure TCP and UDP bandwidth performance.