[ “The distance is nothing; it is only the first step that is difficult.” – Mme Du Deffand ]
In a previous article, Quick Review of IP Routing, we presented the complementary concepts of building the routing table (related to the control plane) and using the table to forward packets (data plane). The current post focuses on the influence of the Administrative Distance (AD) parameter on selecting a route (a control plane task) that will be added to the routing table.
There are multiple routing protocols and each of them uses different units to evaluate how distant a certain destination is. For example, EIGRP costs employs a combination of delay and bandwidth, whereas OSPF bases its calculations solely on bandwidth. RIP, a protocol from the early days of internetworking, selects routes based on a very limited criterion called hop count.
A little reflection on the topic will show that it is not adequate to directly compare routes from different protocols. To somehow take into account the nature of each protocol and to produce a measure of how precise they are, the Administrative Distance concept was introduced. For instance, the default point of view of a Cisco router is to consider EIGRP better than OSPF, which in turn takes precedence over RIP.
Some important aspects to keep in mind to better understand the Administrative Distance:
- AD is only taken into account by a router to compare two equal length network prefixes available to a certain destination. Prefixes of distinct lengths are not comparable. For example, a router that knows the prefix 10.10.10.128/25 via OSPF and 10.10.10.0/24 via EIGRP will install both… (why ?)
- AD is evaluated before any metric information. While AD establishes a comparison between routing protocols, the metric value (or cost) is used to contrast the routes of a certain protocol. This is emphasized by the fact that the AD is the first number inside the brackets that follow the prefix (Figure 1).
In the scenario of Figure 1, the CENTRAL router initially has an OSPF route (through R1) to 192.168.30.0/24. Figure 2, in turn, illustrates what happens when a second route to the same destination gets to be known by CENTRAL.
Figure 2 depicts a situation in which a new router (R2) started advertising an alternate path to the 192.168.30.0/24, originally reachable via OSPF. Upon receipt of this alternate route, the CENTRAL router detects the lower AD value (90 for EIGRP internal versus 110 for OSPF) and the original route is replaced. The information about this new route is also visible in the figure.
- It is important not to confuse the activities of installing the routes in the routing table and using them to forward packets. If you have two routes installed, 10.10.10.0/24 and 10.10.10.128/25, the most specific (for a given destination) will be used. A packet destined to 10.10.10.200/32, for example, will be forwarded using the second route whereas a packet bound to 10.10.10.100 will use the first. ( – Can you explain why ? )
- EIGRP was designed to support a composite metric that includes delay, bandwidth, reliability, load and MTU but, by default, uses only the first two. The EIGRP metric is deemed more precise because it refers to the minimum bandwidth and to the sum of delays along the path from the router to the destination.
- As a reference, the default values for the Administrative Distance parameter for the main types of routes are presented in the following table:
Type of Route
(Default) Administrative Distance
EIGRP External (D EX)
** Related Posts:
- Routing Series: https://alexandremspmoraes.wordpress.com/tag/routing/