Completed "EIGRP Routing: Concepts and Planning" nugget tonight;

Even though Cisco designed EIGRP to be a simple protocol, there are a few concepts you'll want to be aware of before you move forward with implementation. This nugget walks through the key tables, terms, and calculations you'll need to get started with using EIGRP.

 NOTES

WHY USE EIGRP?

1) Backup routes (fast convergence / DUAL)
2) Simple configuration
3) Flexability in summarisation
with OSPF you have to summarise at the ABR's or ASBR's
EIGRP = any point/interface can be do summarisation
4) Unequal cost load-balancing
ONLY protocol to do this (done via variance command)
5) Combines best of distance vector and link state
OSPF is fast, but not as fast as EIGRP due to the Topology table
It is a distance vector protocol (or advanced) because it only knows what it neighbours told it


Router running EIGRP maintains 3 tables;

Neighbour Table
If an adjanacy is formed, the relationship goes in here, done via hello protocol (as opposed to broadcast/multicast like RIP)

Topology Table
Holds the Sucessor routes in here (they are also in the routing table) along with Fessiable successor routes (could be 2 or 3 or so, so long as not "too bad" a route)

Routing Table
Sucessor routes in here along with routing table



EIGRP NEIGHBOURS MSGS

Hello:
Updates (unicast) and Neighbor discovery on multicast 224.0.0.10
NBMA network (i.e virtual circuit, frame relay, ATM etc) hello is every 60 seconds and HOLD time of 180 seconds
Point to Point (other) hello is every 5 seconds and HOLD time of 15 seconds.
Update: 
Once neighbour is formed, router sends FULL routing table and waits for ACK back
Once formed only Hellos will be seen unless route goes down then updates will be sent (msg generally multicast*)
Query:
If network is not available then a DUAL QUERY is sent out for a backup route if one is not in the topology table
Reply:
Are sent in response to query packets, and a reply packet indicates that a new route to the destination has been found.  Update, query and reply packets all use RTP and are considered reliable
Ack:
EIGRP requires acknowledgements from routing updates. (msg considered unreliable)


* During the initial exchange of routes between two new EIGRP neighbors, update packets are unicast rather than multicast

EIGRP = IP Protocol 88

Metric of 47019776 is infinite metric/unreachable, just like hop count of 16 in RIP, this will be used if the neighbour does not have a route to the network


UNDERSTANDING EIGRP METRIC CALCULATION

Bandwidth (K1)
Delay (K3)
Reliabilty (K4 & K5)
Loading (K2)
Does not use MTU (although that is seen)

An EIGRP update to a neighbor contains the cumulative value of each K weight (such as the sum of delays, and lowest bandwidth).  From that, the neighbor receiving the update can calculate the sending neighbors metric for that route


Calculating the EIGRP default metric:

[(10,000,000/lowest Bandwidth) + (sum of delays in 10's of usec)] * 256
10,000,000 is what 10^7 equals

lowest Bandwidth = 10^7/BW, In kbits, BW output from ‘show interface’
Delay in 10′s of usec = DLY output from ‘show interface’

* Note: If the bandwidth division results in a decimal number, round down


So taking a route from my old lab,

Lets take this route that Router A knows about to the network on Router C/D

D       10.14.0.132/30 [90/2174976] via 10.14.0.2, 00:00:58, FastEthernet0/0

2610XM-A#show ip route 10.14.0.132
Routing entry for 10.14.0.132/30
  Known via "eigrp 20", distance 90, metric 2174976, type internal
  Redistributing via eigrp 20
  Last update from 10.14.0.2 on FastEthernet0/0, 00:45:25 ago
  Routing Descriptor Blocks:
  * 10.14.0.2, from 10.14.0.2, 00:45:25 ago, via FastEthernet0/0
      Route metric is 2174976, traffic share count is 1
      Total delay is 20200 microseconds, minimum bandwidth is 1544 Kbit
      Reliability 255/255, minimum MTU 1500 bytes
      Loading 1/255, Hops 2

2610XM-A#show ip eigrp topology 10.14.0.132 255.255.255.252
IP-EIGRP (AS 20): Topology entry for 10.14.0.132/30
  State is Passive, Query origin flag is 1, 1 Successor(s), FD is 2174976
  Routing Descriptor Blocks:
  10.14.0.2 (FastEthernet0/0), from 10.14.0.2, Send flag is 0x0
      Composite metric is (2174976/2172416), Route is Internal
      Vector metric:
        Minimum bandwidth is 1544 Kbit
        Total delay is 20200 microseconds
        Reliability is 255/255
        Load is 1/255
        Minimum MTU is 1500
        Hop count is 2

We can see the following information for this route 10.14.0.132/30
Route metric is 2174976 
Total delay is 20200 microseconds
Minimum bandwidth is 1544 Kbit

So applying the above calculations to work out the metric;

10^7 = 10000000 ÷ 1544kbits (min bandwidth) = 6476.6839.... (round down to 6476)
6476 + all delay links which is 20200ms (20200ms ÷ 10 = 2020usec) = 8496
8496 X 256 = 2174976 is the metric for that route :0)

GOD I HATE MATHS!!! but there you go, i think cisco thought hang on a minute, we have a super kool routing protocol but we have made it tooo easy, lets complicate this a little .... jesus!

Right, over the next couple of days lets start planning/building my new lab topology ;0)