Tips on Balancing Your Server for Network Applications

High-volume network applications are driving the need for optimization technologies. Web transaction rates and throughput continue to increase exponentially with Internet-driven applications and media-rich communications. In response, customers in all verticals are considering deploying or upgrading server load-balancing technology to handle high-volume network applications.

Here are four tips for choosing the right server load-balancing technology for your Web applications.

1. Choose the right platform. In order to address the need for higher online transmission speeds, vendors continue to move their single-CPU software solutions to more powerful multi-CPU load-balancer systems. To future-proof a load-balancer upgrade or investment, determine whether the vendor's platform is optimized for multicore CPU systems. A properly designed multi-CPU system will have maximum per-CPU independence with almost no interference among CPUs, making instructions run faster.

A key benefit of the multicore CPU platform is that the server load balancer performs as desired with numerous features turned on, such as Layer 4, Layer 7, secure socket layer (SSL) and advanced scripting features, which are CPU intensive. The advanced scripting features allow administrators to solve traffic bottlenecks before they occur. A recommended way to test the vendor's platform is to deploy them into the network and test the performance while running the scripting feature to make sure degradation is not significant.

Also, be careful with solutions that are getting most of their Layer 4-7 acceleration functions from hard-wired application-specific integrated circuit (ASIC) technologies. Getting simple feature requests into the accelerating ASIC requires 18 months or more development time.

2. Acceleration, optimization and security. Server load balancers should have the architecture to accelerate and improve Web application response times. When tuned properly, customers can benefit from Web site and application response times up to 10 times faster.

The solution should have the intelligence to optimize and reduce bandwidth, space and energy requirements. Customers' bandwidth savings can...