Three server load balancing user classification

Users who use server load balancing can be classified into the following three categories according to simple classification:

The first category, performance requirements.

The traffic has soared, and a single server can't stand it. The clients here include all kinds of websites, whether it is content, community, marriage, direct sales, whether you are selling clothes, selling bags or selling tea, the traffic growth caused by the increase in the number of users, the use of load Balance is a matter of time. And for the majority of website customers, usually the simple 4 layer traffic sharing can not meet their needs. Memory caching, hardware compression, connection multiplexing, etc. ADC functions that can reduce server load are becoming more and more a must for website customers. Another point is that the application of the website customer is complicated. There are often multiple domain names corresponding to the same IP and the same server responding to multiple different domain names. According to the host (domain name) switch, even the 7-layer traffic distribution script requirements are very common. demand.

Another major category of performance pressure is the various operators. Each operator's online business hall, internal DNS system, Radius user authentication system, almost must configure server load balancing. However, operators use load balancing in a wide range of areas. Several key systems are due to performance pressures. Most other applications may be more due to other considerations.

The second category, the expansion requirements.

This category is sometimes difficult to distinguish strictly from performance requirements. The most common is the need to consider load balancing due to bottlenecks caused by software platforms or software development. For example, an enterprise user who uses load balancing very early is not because the hardware or bandwidth requirements of the server are not satisfied, but the software performance limitation of the important internal ERP system based on software development. A single server cannot meet internal requirements at all. Therefore, it is necessary to use load balancing to expand.

The typical users here are the internal systems of large enterprises. Many enterprises have to make many internal applications tailored to their own characteristics due to their internal needs. These applications are usually handed over to internal or outsourced development, with a focus on the system. Functional integrity is usually weaker for performance considerations. It is often easy to expose potential performance defects when used on a large scale and must be solved with simple extensions. Some applications of operators also belong to this category.

An obvious feature of this type of application is that there are still many applications that use the C/S architecture and use their own unique client for TCP communication. When load balancing is deployed, it is usually necessary to consider changing the network structure of the client. Because either the server needs to record the client source address, or the server has a problem with multiple request processing initiated by the same address of the NAT class.

The third category, reliability requirements.

This category has relatively low performance requirements. For example, on government websites, traffic is usually not high, but requires relatively high reliable online time. Four 9 (ie 99.99% reliability) are converted to a failure time of 4 months per month, no more than one hour a year. It is almost impossible to achieve this standard by hand or cold. Load balancing, never crash. For such needs, load balancing devices have become a must-have for securing SLAs (Service Level Agreements). It should be noted that in this type of deployment, including load balancing itself, it also needs to provide dual-system HA hot standby and the mechanism for user connection information backup. In the case of single load balancing itself, the entire system can still be guaranteed due to hardware and software failures. Reliable operation.

The above is a summary of the simple use of server load balancing customers, just three functions that can be provided for server load balancing, namely Performance Scalability Availability, high performance, scalability and reliability. In fact, in many cases, the three are related to each other. It is not uncommon for customers to deploy load balancing only for one reason.