How to solve the high load of the Hong Kong server? Renting the Hong Kong server deployment website. During the launch of the website business, it will inevitably encounter a high traffic access. When this happens, Hong Kong servers often face high loads. If the server load is too high, if it is not solved in time, it will often cause the server to crash.

　　How to solve the high server load in Hong Kong? At this time, the server load balancing is needed.

　　1. Cluster: Use DNS to implement the server load balancing, that is, set multiple A records on the DNS server. The principle is that different DNS requests will be parsed to different IP addresses. This is like a summary website for various classification processing data In the same way, each category has different request addresses. After processing, return to the user together, thereby sharing the load of the main server. This is often the cluster technology. The disadvantage is: DNS takes effect for a long time and poor scalability.

　　2. IP -based load balancing: The most representative belongs to LVS. The principle is to obtain IP information through the Linux kernel, and then forward the IP information to a server in the cluster according to the specified load balancing algorithm. Access traffic or data is processed on average to multiple node devices, which can reduce the response time of the webpage. The disadvantage is that the performance of LVS is more dependent on the Linux kernel, which will lead to the reduction of LVS performance in the long run.

　　3. Google MAGLEV: Maglev to load balancing by Google, you can process 100W requests per second without preheating. Under the 8 -core CPU run by the MAGLEV instance tested by Google, the upper limit of the network throughput reaches 12M PPS per second. If Maglev uses the Linux kernel, the speed will be less than 4m PPS per second.