Flooding

Flooding

Flooding is a reliable method to distribute information within a network.

Its main advantage, but also its main drawback is redundancy. On receipt of a PTSE that is not yet in its database, a node forwards this PTSE to all neighbors , except the one the PTSE was received from. If there is more than one path between any two nodes, a PTSE will be forwarded over each of them. Therefore, appropriate measures have to be taken, to prevent redundant PTSEs from consuming to much processing power at the receiving node.

Each received PTSE is checked, whether it is already installed in the nodal database. Following, there are two actions possible:

– Discarding the PTSE, if it is already installed in the database.

– Forwarding the PTSE via flooding, installing it in the database and then acknowledging of the PTSE.

There are two major reasons, why a node originally floods a PTSE:

–Triggered Update: Triggered flooding happens if a completely new PTSE

is originated by a node or if there is a significant change in an IG within an existing PTSE (e.g. new end system addresses are added, the availablebandwidth changed beyond a threshold etc.).

–Aging: Aging causes flooding if either the remaining lifetime of a PTSE

reaches zero or if the remaining lifetime of the PTSE reached a certain threshold in its originating node. To prevent the PTSE from being deleted the originating node floods an update, even if the contents did not change. Summarizing, while database synchronization is limited to the moment where two neighbors learn about their existence, flooding lasts as long as the
network is up and running.