Network load balancing is a method of allocating tasks across multiple network devices (e.g. servers) in order to optimize resources usage, reduce the request processing time, as well as to provide fault tolerance (redundancy).
Load balancing can be used to empower the server farm. It also allows to continue server operation even when several execution units are out of order.
To achieve maximum bandwidth and resilience you have to keep several important rules.
- Reliably protect NLB nodes and load balancing applications. The NLB subnet should be physically protected from intrusion of unauthorized users and devices in order to avoid cracking.
- Use only TCP/IP network protocol on the cluster adapter. In fact, adapter needs no other protocols.
- Use Network Load Balancing Manager as it can significantly quicken the process of configuring. Note: the simultaneous use of NLB Manager and Network Connections folder to change properties may lead to unpredictable results.
- Disable NLB remote control as it exposes the security system to unnecessary risk (e.g. data corruption, failure of services, unauthorized access, and so on). It is strongly recommended not to enable remote control. Use NLB Manager or other remote management tools instead.
- Enable NLB Manager logging. Such log can be very useful in troubleshooting. Note: the NLB Manager log file contains potentially important information about the cluster nodes and network load balancing, so it must be reliably protected.
- Make sure that any load balancing application is running on all cluster nodes it is installed on.
It should be noted in conclusion that this method is easily extensible: just add one more server and specify it in the NLB cluster. The only restriction is that NLB supports up to 32 nodes.