“If the purpose of an IP network is to deliver traffic from Point A to Point B, then routing is simply the process of moving traffic from A across a particular path through multiple links and routers to get to B.”

So say officials of Packet Design (News - Alert), adding that while that process might seem simple enough, “the choice of which path to use for any given set of traffic transiting the network is not so simple.” That’s where routing protocols come into play.

Netflow protocols, for those unfamiliar with them, are the combination of software and signals “used to calculate and communicate the best network paths for successful traffic delivery,” according to company officials, who explain that the software runs on all the routers in the network, and the signals are passed between all the routers on the network.

“Of course, this process doesn’t just happen once,” company officials note. “Every time things change, which they often do, the process repeats to keep all routers abreast of the best traffic paths and thereby allow the network to fulfill its purpose.”

They give some examples of the changes that cause the routing protocol process to update itself across the network:

Loss of a network link.

This could be an Ethernet network, for example. It causes all the network addresses accessible via that link to become unreachable.

Loss of a link between routers.

This causes previously valid paths to become invalid, requiring a recalculation of traffic paths.

Then there’s the addition of new network links or network addresses, which would require recalculation of best paths.

Or there could be a soft or hard reset of a router, or the downing of a router for maintenance or due to a failure. This affects multiple network links, network addresses and paths.