In this example a single BalanceNG node is connected with one interface (“leg”) to one local Layer 2 network (LAN), which transports two Layer 3 IPv4 networks. The load balancing is configured between those two Layer 3 networks towards two target servers. It may be well used for getting started with BalanceNG concepts, but running two Layer 3 networks on a single Ethernet LAN may be confusing in the beginning. One advantage is, that there’s no need to change network topology like cables and switches. Runs with BalanceNG V3 and higher.
This example is similar to the Single Legged Routing Mode Example, but connecting two different LANs (one for the targets / real servers and the other for representing virtual servers). Additionally it implements high availability with VRRP. This topology setup is often used with Telco applications.
This example shows how to setup a so called “Direct Server Return” (DSR) network topology. In general, DSR comes with important advantages: It’s usually the fastest, quite easy to implement and it requires no network changes or additional components. One thing that doesn’t work with DSR is port rewriting. And: Beware of the ARP flux 🙂 !
In LLB (Link Load Balancing) mode BalanceNG is configured to perform as a router balancing the bandwidth of an office network (for example) to different physical uplinks. As a side effect IP connectivity becomes high available.
This example shows how to implement a virtual cross-cable which connects two virtual TAP interfaces (Linux only). This setup is implemented by one instance, where the other instance may be used to examine the HW performance using the BalanceNG benchmark command.