How many reasons are there for high network latency? How to solve?

1. Too many routing hops between the local machine and the server. Since the transmission speed of light/electricity is very fast, their propagation time in the physical medium is almost negligible, but the processing time of the router to forward the packet is non-negligible. Network latency is noticeable when there are too many routes in the local-to-server link to forward processing.

2, not enough network bandwidth. Excluding other factors, if the client and server side directly through a router connection, but the bandwidth is only 10Kbps, but at the same time, there are multiple applications need to transmit far more than the bandwidth of the amount of data 200Kbps, which will cause a lot of data loss, which is manifested in the response delay.

3, processing bandwidth is not enough. Excluding other factors, if the client and server-side direct connection through a router, and the bandwidth is sufficient, but the server-side processing capacity is insufficient, will also cause response delays.

Solution:

1. Improve WAN performance

Enterprises can meticulously control the performance of applications within the LAN, but that control does not extend to the WAN. a WAN typically has multiple optional service providers who operate carrier-grade, top-of-the-line backbone infrastructures. WAN operators can also contribute to reducing latency by choosing shorter and more efficient routing paths, deploying low-latency switches and routing equipment, and proactively avoiding network equipment downtime.

Increasing WAN bandwidth improves application performance, but bandwidth isn't cheap and it usually doesn't have to be. In practice, WAN application performance can also be improved by employing a variety of techniques that make more efficient use of available WAN bandwidth. These techniques are collectively referred to as WAN gas pedals. Gas pedals function by reducing the data payload and utilizing the available WAN bandwidth more efficiently.

2. Fixing application performance on the LAN

If an application that has just been installed or whose settings have been modified is experiencing local network performance problems, try verifying the application's settings, system compatibility, and software status; also review the installation and setup documentation. Hardware compatibility can also affect LAN efficiency. In some cases, updates or patches may reverse otherwise poor performance.

3. Workload balancing, moving virtual machines between servers, optimizes application workloads and bandwidth requirements for each physical host.

Data centers might also consider swapping out servers' NICs for 10-gigabit or faster models, or installing separate dedicated network adapters for struggling high-bandwidth applications. Of course, faster NICs can be very expensive, the process of installing the physical components requires the servers to be taken offline, and usually adds to the cost of the supporting LAN switching infrastructure.

Expanded Information

Network Latency Levels: (The lower the network latency PING value, the faster the speed)

1~30ms: Very fast, almost imperceptible latency, playing any game at a particularly smooth speed;

31~50ms: Good, can be a normal game, there is no noticeable delay;

51~100ms: ordinary, confrontational games at a certain level or more can feel the delay, occasionally feel the stutter;

100ms~200ms: poor, can not play confrontational games, there is obvious lag, occasionally appear packet loss and drop phenomenon;

200ms~500ms: very poor, access to the web has a

200ms~500ms: very poor, access to the web page has obvious delay and lag, often packet loss or inaccessible;

>500ms: very poor, unacceptable delay and packet loss, or even unable to access the web page;

>1000ms: basically inaccessible;

Calculation: 1 second = 1,000 milliseconds (Example: 30ms is 0.03 seconds).

Baidu Encyclopedia-Network Latency