Chapter_09_2

EIGRP keywords : distance vector (or hybrid), classless, RTP, bounded updates, DUAL Algorith, Adjacencies, Neighbor Table, Table

EIGRP Packet Types

EIGRP uses five different packet types, some in pairs.

Hello packets are used by EIGRP to discover neighbors and to form adjacencies with those neighbors. EIGRP hello packets are multicasts and use unreliable delivery. EIGRP Hello packets are discussed in a later section.

.

Update packets are used by EIGRP to propagate routing information. Unlike RIP, EIGRP does not send periodic updates. Update packets are sent only when necessary. EIGRP updates contain only the routing information needed and are sent only to those routers that require it. EIGRP update packets use reliable delivery. Update packets are sent as a multicast when required by multiple routers, or as a unicast when required by only a single router. In the figure, because the links are point-to-point, the updates are sent as unicasts.

Acknowledgement (ACK) packets are sent by EIGRP when reliable delivery is used. RTP uses reliable delivery for EIGRP update, query, and reply packets. EIGRP acknowledgement packets contain a nonzero acknowledgment number and always are sent by using a unicast address.

Query and reply packets are used by DUAL when searching for networks and other tasks. Queries and replies use reliable delivery. Queries use multicast, whereas replies are always sent as unicast. DUAL is discussed in a later section. Query and reply packets are discussed in more detail in CCNP.

All neighbors must send a reply regardless of whether or not they have a route to the downed network. Because replies also use reliable delivery, routers such as R2, must send an acknowledgement.

Note: You may be wondering why R2 would send out a query for a network it knows is down. Actually, only the interface attached to the network is down. Another router could be attached to the same LAN. Therefore, R2 queries for such a router before completely removing the network from its database.

On most networks EIGRP Hello packets are sent every 5 seconds. On multipoint nonbroadcast multiaccess networks (NBMA) such as X.25, Frame Relay, and ATM interfaces with access links of T1 (1.544 Mbps) or slower, Hellos are unicast every 60 seconds. An EIGRP router assumes that as long as it is receiving Hello packets from a neighbor, the neighbor and its routes remain viable.

Holdtime tells the router the maximum time the router should wait to receive the next Hello before declaring that neighbor as unreachable. By default, the hold time is three times the Hello interval, or 15 seconds on most networks and 180 seconds on low speed NBMA networks. If the hold time expires, EIGRP will declare the route as down and DUAL will search for a new path by sending out queries.

EIGRP uses the term partial or bounded when referring to its update packets. Unlike RIP, EIGRP does not send periodic updates. Instead, EIGRP sends its updates only when the metric for a route changes.

The term partial means that the update only includes information about the route changes. EIGRP sends these incremental updates when the state of a destination changes, instead of sending the entire contents of the routing table.

The term bounded refers to the propagation of partial updates sent only to those routers that are affected by the change. The partial update is automatically "bounded" so that only those routers that need the information are updated.

By sending only the routing information that is needed and only to those routers that need it, EIGRP minimizes the bandwidth required to send EIGRP packets.

The router eigrp autonomous-system global configuration command enables EIGRP. The autonomous system parameter is a number chosen by the network administrator between 1 and 65535. The number chosen is the process ID number and is important because all routers in this EIGRP routing domain must use the same process ID number (autonomous-system number).

The metric for EIGRP is calculated by EIGRP Metric = 256*([K₁*Bw + K₂*Bw/(256-Load) + K₃*Delay]*[K₅/(Reliability + K₄)]).

or Metric = [ (10.000.000 / minimum bandwidth ) + (sum of delay/10 ) ] * 256

DUAL uses several terms which will be discussed in more detail throughout this section:

Successor

Feasible Distance (FD)

Feasible Successor (FS)

Reported Distance (RD) or Advertised Distance (AD)

Feasible Condition or Feasibility Condition (FC)

These terms and concepts are at the center of DUAL's loop avoidance mechanism. Let's examine them in more depth.

A successor is a neighboring router that is used for packet forwarding and is the least-cost route to the destination network. The IP address of a successor is shown in a routing table entry right after the word via.

Feasible distance (FD) is the lowest calculated metric to reach the destination network. FD is the metric listed in the routing table entry as the second number inside the brackets. As with other routing protocols this is also known as the metric for the route.

The successor, feasible distance, and any feasible successors with their reported distances are kept by a router in its EIGRP topology table or topology database. As shown in the figure, the topology table can be viewed using the show ip eigrp topology command. The topology table lists all successors and feasible successors that DUAL has calculated to destination networks.

R2#show ip eigrp topology 192.168.1.0

By default, EIGRP will use only up to 50 percent of the bandwidth of an interface for EIGRP information. This prevents the EIGRP process from over-utilizing a link and not allowing enough bandwidth for the routing of normal traffic. The ip bandwidth-percent eigrp command can be used to configure the percentage of bandwidth that may be used by EIGRP on an interface.

Router(config-if)#ip bandwidth-percent eigrp as-number percent