The application of load balancing technology in campus network based on LVS-NAT

Abstract. The concurrent processing and load capacity of a single server cannot meet the growing demand of users for a variety of services in a campus network system. This document put forward to solve this problem using load balancing techniques based on LVS-NAT, discussed the key technologies of LVS-NAT, designed and implemented campus network service system with LVS-NAT load balancing technology and tested. The results showed that this system improved the processing and load capacity of the concurrent server effectively and provided a good reference to building the efficient and stable digital campus network system.

Keywords: load balancing, campus network, LVS-NAT, cluster system

1. Introduction

With the continuous development of information technology, people pay more and more attention to digital campus’s building. In most of the campus network system, concurrent processing and load capacity of a single server cannot meet the growing demand of users for a variety of services. Therefore, the primary task is to improve the performance of the server. Re-purchasing a new higher performance server is not only costly but also wasteful of resources. So the paper introduces the LVS-NAT technology. This technology not only extends the processing capacity of a single server in campus network to multiple peer-to-peer servers by the cluster technique, but also distributes client’s requests to a server array averagely through the load balancing strategy[1]. It provides fast access to important data and solves a large number of concurrent access services[2].

2. LVS-NAT load balancing technology

LVS (Linux virtual server) is a virtual server cluster system., composed of multiple real servers, which are transparent to the users[3]. The virtual server receives IP packets from the users, sends those packets to the real server in the cluster system by using some scheduling policies. The real server handles user’s requests. LVS-NAT means that the Linux virtual server sends user’s packets to the real server by network address translation technology which is called NAT for short. LVS-NAT system mainly includes the following aspects[4]:

2.1 Load Balancer

When users access network services through the virtual IP address, user’s request packets reach the load balancer firstly, the load balancer selects a server from a set of real servers based on connection scheduling algorithm[5]. The load balancer rewrites the target address and port of user’s request packets into the selected server’s address and port, finally it sends the revised packets to the selected server. After the request is completed by the server, it will be returned by the load balancer to the user.

2.2 Server Pool

It is a cluster system composed of some real servers providing the services. The number of the real servers in this system can be extended dynamically. You can configure the Web service, Email service, FTP services and other services on the servers. The load balancer selects a certain server to provide service according to the network environment[6].

2.3 Backend Storage

It is a shared storage area for the server pool. The server pool has the same content, providing the same service through accessing the backend storage[7,8].

3. The implementation of LVS-NAT technology in the campus network

3.1 The framework of the system

Through the analysis of research on LVS-NAT technology and the demand of campus network services, the overall framework of this system consists of four layers as follows:

figure1. The framework of the system

The first layer is the client. It includes students、teachers who send requests through the campus network. The second layer is the Linux virtual server which is the entrance of campus network. It is responsible for receiving the client’s requests, monitoring the working states of each server in the server pool， resending requests to the real server by using NAT technology and returning the service messages to the first layer. The third layer is the server pool. It is a cluster system composed by some real servers. The fourth layer is a shared area provided for the real server.

3.2 The implementation of the system

The clients connect to the load balancer through the campus network. The core of this system is the configuration of load balancing server and the server pool.

3.2.1 The Selection and configuration of configuration

First, choose a server with good performance and install Linux operating system on it as load balancing server, Load balancing server has two network cards which connect the external network and internal network respectively. Set up the backup servers for the load balancing server and do the same configuration on them.

Second, install the software of Ipvsadm in load balancing server. The software is in the kernel of Linux used for monitoring IP packets from the user, processing and resending packets to the real server.

Third, install the heartbeat software in the two load balancer .Heartbeat monitors the running status of load balancer by serial device. Each load balancing server runs a daemon process, known as the "heart" program to read and write other heartbeat’s child process in other load balancer in order to know the state of other load balancer.