[c-nsp] Unintentional load balancing of traffic

[Mark Tinka]

On 18/Sep/15 10:43, Raymond Burkholder wrote:

> Hazarding a guess:  you have explicit costs assigned to the local links of
> 201 and 202.  But as OSPF is a link vector protocol, all costs between local
> interface and destination interface are added together.  So when you add the
> cost of all links between source and destination, they must add up to 205 on
> each of two paths.  And thus get load balanced.

A reasonable theory.

Just to clarify, OSPF is a link state routing protocol :-). But yes, in
essence, as with all routing protocols, the end-goal is to calculate the
best vector toward another node.

Mark.