WEB Server Load Balancing System Design And Implementation

The rapid growth of the Internet makes web applications become the most important and most widely used. But with a high-performance computers has been difficult to meet the growing demand for services. Load balancing on the existing network architecture, it is possible to provide an inexpensive and effective way to extend the network bandwidth, network congestion resolved, the server provides the free response, improve user response speed, to avoid a single point of failure.Various information systems as a core part of the volume of business increased, the rapid growth of traffic and data traffic, and its processing capacity and computational intensity will also increase accordingly, so that a single device simply can not afford to be using multiple servers to work together to improve computer processing and computational intensity systems to meet the current needs of the business volume.In addition to critical load balancing system load balancing algorithm outside, other modules are also designed to determine the level of system performance. For example, the system is running, creating and destroying objects is a very timeconsuming, so a means to improve the efficiency of the system is to minimize the number of times to create and destroy objects, because this is some of the " pooled resources " technology produced. This paper analyzes the load balancing principle, based on the theoretical analysis, the proposed method combines the thread pool, the memory pool, slab caches, such as load balancing pool technology systems.The basic realization of the functional requirements of the system, and in the case of load balancing and web cache performance than ordinary service cluster will be greatly improved. Can provide up to more than ten kinds of load balancing scheduling algorithms, including : rotation, weighted rotation, minimum connection, weighted least connection, random, weighted random, based on source / destination address / source address port HASH, local priorities, proximity selection, scheduling and dispatching based bandwidth HTTP content-based algorithm. Fully functional and flexible deployment location to meet the needs of a variety of load balancing applications.