Install guide

1-2 Release Note
Software Version 2.7.5
C613-10454-00 REV A
Introduction
This chapter describes the WAN load balancing feature, how it is supported on
the router, and how you can configure its operation.
With the increasing use of the Internet to service core business functions comes
the need for reliable WAN connectivity. A specific aspect of this requirement is
the need for reliable connectivity to specific destinations. This requirement can
be simply and effectively met by providing alternative network connections
via different Internet service providers (ISPs). In this way an outage limited to
one ISP will not result in a loss of connectivity to remote destinations,
providing these are still accessible via the other ISP.
Tip For information about other methods of load balancing, see the Server
Load Balancing chapter of your router’s Software Reference.
Operating Principles
When a WAN load balancing router simultaneously connects to multiple WAN
networks, it will try to distribute its traffic equally across each network
interface. A typical example is a router that has two Internet connections, each
exchanging data to remote sites via different Internet service providers (ISPs).
In this case you can configure the load balancer to balance its traffic based
either on the traffic profile to each port’s ISP, or to specific remote destinations.
Although connectivity via multiple WAN interfaces can be achieved using
routing protocols such as RIP and OSPF; these protocols usually choose their
routing paths based on routing metrics rather than on dynamic load
conditions. For example, if a router has two WAN ports and each port connects
to a different ISP, the router will send most of its traffic via the port offering the
best metric. Although this method provides alternative connectivity in the
event of an ISP network failure, under normal operating conditions it wastes
the bandwidth available via the alternative port.
When a router receives a packet from one of its interfaces, it creates an IP
session (termed a flow) based on the following fields:
■ source and destination IP addresses
■ upper layer protocol used.
When WAN load balancing is enabled, the router creates each load balancer
session based the particular combination of values contained within these
fields. Each field combination is represented by a particular IP flow. The router
then creates a mapping between the particular IP flow and its load balancer
session. The IP flows and the load balancer sessions have a many-to-one
relationship. This means many IP flows can be mapped onto a single load
balancer session. Once the load balancer has applied its algorithm to determine
the best balanced route to use, it remembers this route for future traffic.
Therefore IP flows that share the same LB session will use the same route for
forwarding.