Chapter 37 Ospf; Introduction To Ospf - Planet XGS3-42000R User Manual

37.1 Introduction to OSPF

OSPF is abbreviation for Open Shortest Path First. It is an interior dynamic routing protocol for autonomous
system based on link-state. The prot ocol creates a link-state database by exchanging link-states among
layer3 switches, and then uses the Shortest Path First algorithm to generate a route table basing on that
database.
Autonomous system (AS) is a self-managed interc onnected network. In large networks, such as the Internet,
a giant interconnected network is broken down to autonomous systems. Big enterprise net works connecting
to the Internet are independent AS, since the other hosts on the Internet are not managed by those AS and
they don't share interior routing information with the layer3 switches on the Internet.
Each link-state Layer3 switch can provide information about the topology with its neighboring Layer3
switches.
• The network segment (link) connecting to the lay er3 switch
• State of the connecting link
Link-state information is flooded throughout the network so that all Layer3 switches can get firsthand
information. Link-state Layer3 switches will not broadcast all information contained in their route tables;
instead, they only send changed link-state information. Link-state Layer3 switches establish neighborhood by
sending "HE LLO" to their neighbors, then link-state advertisements (LSA) will be sent among neighboring
Layer3 switches. Neighboring Layer3 switch copy the LSA to their routing table and trans fer the information to
the rest part of the network. This process is referred to as "flooding". In this way, firsthand information is sent
throughout the network to provide accurate map for creating and updating routes in the network. Link-state
routing protocols use cost instead of hops to decide the route. Cost is assigned automatically or manually.
According to t he algorithm in link-state protocol, cost can be used to calculate t he hop number for packets to
pass, link bandwidth, and current load of the link. The administrator can even add weight for better
assessment of the link-state.
1) When a link-state layer3 switch enters a link-state interconnected net work, it sends a HELLO packet to
get to know its neighbors and establis h neighborhood.
2) The neighbors respond with information about the links they are connecting and the related costs.
3) The originate layer3 switch uses this information to build its own routing table
4) Then, as part of the regular update, layer3 switch send link-state advertisement (LSA) packets to its
neighboring layer3 switches. The LSA include links and related costs of that layer3 switch.
5) Each neighboring lay er3 switch copies the LSA packet and passes it to the next neighbor (i.e. flooding).
6) Since routing database is not recalculated before layer3 switch forwards LSA flooding, the converging
time is greatly reduced.
One major advantage of link-state routing protocols is the fact that infinite counting is impossible, this is
because of the way link-state routing protoc ols build up their routing table. The second advantage is that
converging in a link-state interc onnected net work is very fast, once the routing topology changes, updates will
be flooded throughout the net work very soon. Those advantages release some lay er3 switch resources, as
the process ability and bandwidth used by bad rout e information are minor.

Chapter 37 OSPF

37-1