With the increasing popularity of Internet, the users and the business volume have grown exponentially, and the users have stronger and stronger dependence on a variety of Internet-based key businesses, while traditional server deployment architecture can not really solve the users' problems in terms of performance, scalability, security and so on. 

Service-end performance bottleneck

When the users face the single point of failure and server performance bottlenecks, they usually solve them by adopting one of the two ways below:
First, replace the existing server by the server with greater processing power, so as to improve performance; improve the availability by means of duplex redundancy.
Second, build server clusters by increasing the number of servers. 

Both of these methods have limitations. The first solution is more expensive, and does not have very good scalability; service has to be interrupted when maintenance and upgrade is made, and the cost of replacing servers includes the cost of new server, moreover, the original server could no longer play a role though it is still in a well-run state. The second solution is a relatively cheap solution which can improve the website's processing capacity by adding new servers, and make new servers work with the original server, so as not to interrupt service when upgrade is made; however, the weakness of the second solution is that each server has a unique IP address, users need to remember multiple IP addresses to better access to the site, which also results in ineffective distribution of traffic between multiple servers.

Bandwidth bottleneck

When the users face the bandwidth bottleneck, they usually solve them by increasing the bandwidth, but the increase in bandwidth means increased costs, moreover, the speed of bandwidth increase will never keep up with the speed of bandwidth being occupied. Individual users or individual applications can easily take up a lot of bandwidth resources, and makes the traffic of P2P, streaming media which are unrelated to the said service may take up most of the bandwidth, so the unrestricted increase in bandwidth is not realistic. 

In such cases, users need a solution to really offer 7*24 high-reliability guarantee, and enhance performance and security of key businesses. For this reason, Wafer recommends network intelligent optimization solutions based on Radware’s AppDirector + AppXcel, Bluecoat’s ProxySG, and Packeteer’s PacketShaper.