|
0:00:14
|
In our next section here, we're gonna
|
|
0:00:19
|
which we'll see it has more features and
|
|
0:00:22
|
some more design understanding that
|
|
0:00:27
|
But it's still not gonna be as complicated as
|
|
0:00:29
|
the OSPF or BGP features and design issues that
|
|
0:00:38
|
So, we'll see when we look at both
|
|
0:00:44
|
as compared to OSPF and BGP, there's
|
|
0:00:49
|
test out before getting into the exam.
|
|
0:00:51
|
But just like RIP, you're pretty much
|
|
0:00:56
|
in the routing and switching lab exam.
|
|
0:00:58
|
So, there's no reason to go to the exam without making
|
|
0:01:05
|
Now, as we know, EIGRP is considered a
|
|
0:01:10
|
because it has some properties of link state protocols
|
|
0:01:16
|
So, as a link state property,
|
|
0:01:21
|
But as distance vector, it still uses things
|
|
0:01:26
|
Now, we'll see that the diffusing update
|
|
0:01:32
|
But technically, the protocol
|
|
0:01:35
|
because we don't have an overall view
|
|
0:01:41
|
So, just like RIP, EIGRP is considered
|
|
0:01:46
|
Which means it only knows what is directly
|
|
0:01:51
|
Now, based on the feasibility condition
|
|
0:01:55
|
we'll see that EIGRP is
|
|
0:01:59
|
and it's highly, highly unlikely that a
|
|
0:02:03
|
But technically, it is still considered
|
|
0:02:08
|
Now, EIGRP like OSPF, does use
|
|
0:02:14
|
which is IP protocol number 88,
|
|
0:02:20
|
So, this means that when we are establishing
|
|
0:02:26
|
we need to make sure that protocol
|
|
0:02:31
|
This will be both for
|
|
0:02:35
|
By deafult, EIGRP uses multicasts to establish
|
|
0:02:42
|
But beyond that, most of the
|
|
0:02:46
|
in order to actually synchronize the topology.
|
|
0:02:50
|
Once the network is fully converged,
|
|
0:02:56
|
but unicast are still used for things
|
|
0:02:58
|
and the query reply messages when
|
|
0:03:05
|
So, if we were to configure an
|
|
0:03:11
|
we would need to take into account both
|
|
0:03:16
|
and the unicast addresses
|
|
0:03:21
|
Now, typically, when we're
|
|
0:03:24
|
if we want to permit it when
|
|
0:03:27
|
maybe the content-based access
|
|
0:03:33
|
The easiest way would just be to
|
|
0:03:38
|
So, if we said, Access List
|
|
0:03:43
|
or Permit 88 from this particular
|
|
0:03:46
|
that 88 is referncing the
|
|
0:03:51
|
When we actually looked at this
|
|
0:03:53
|
the shortcut keyword EIGRP is
|
|
0:03:59
|
Likewise, if we were to say,
|
|
0:04:02
|
it's talking about protocol number 89.
|
|
0:04:08
|
Now, the actual configuration of
|
|
0:04:12
|
when we compare it to OSPF or BGP.
|
|
0:04:15
|
Where we simply enable
|
|
0:04:18
|
and specify what interfaces we
|
|
0:04:24
|
Now, EIGRP can run multiple
|
|
0:04:29
|
which means that the autonomous system
|
|
0:04:33
|
Or as an OSPF, the process ID is only
|
|
0:04:39
|
And EIGRP, the AS number is
|
|
0:04:44
|
So, everyone has to agree on the
|
|
0:04:48
|
and in order to exchange the updates.
|
|
0:04:52
|
Now, the network statement in EIGRP is a little bit more
|
|
0:04:59
|
Because we can control with
|
|
0:05:02
|
exactly what interfaces the
|
|
0:05:07
|
based on the IP address of the link.
|
|
0:05:09
|
So, we'll see, this is similar to to how a
|
|
0:05:14
|
Part of the confusion of this though,
|
|
0:05:16
|
is that the IOS parcer has some error
|
|
0:05:22
|
that if you enter wildcard
|
|
0:05:25
|
as opposed to a wildcard mask,
|
|
0:05:27
|
the parceer will automatically convert
|
|
0:05:32
|
So, if I have an address on the link,
|
|
0:05:35
|
let's say for example that
|
|
0:05:39
|
for the Fast Ethernet 0/1,
|
|
0:05:42
|
we have 155.10.5.0/24,
|
|
0:05:47
|
If I put a network statement under EIGRP,
|
|
0:05:51
|
that says, Network 150.10.5.0 255.255.255.0,
|
|
0:05:57
|
the parcer will take that command,
|
|
0:05:59
|
but then coinvert the subnet
|
|
0:06:04
|
The most accurate way to
|
|
0:06:08
|
would be to match on the
|
|
0:06:12
|
where we would say,
|
|
0:06:26
|
This essentially means that
|
|
0:06:29
|
on any interface that has the
|
|
0:06:36
|
We don't care what the subnet mask is,
|
|
0:06:39
|
we don't care if this is an address that
|
|
0:06:42
|
like in an unnumbered situation.
|
|
0:06:45
|
When we look at the Show IP Interface Brief,
|
|
0:06:48
|
any interface that matches this address
|
|
0:06:55
|
So, part of the confusion here
|
|
0:06:57
|
is that it does not relate to
|
|
0:07:02
|
So, I'm saying, "0.0.0.0" in the wildcard mask,
|
|
0:07:04
|
it does not mean that
|
|
0:07:08
|
It simply means that this interface
|
|
0:07:13
|
Now, we'll see in some
|
|
0:07:17
|
like OSPFv3 and EIGRPv6 for IPv6 routing,
|
|
0:07:21
|
they removed this confusion
|
|
0:07:25
|
by moving the enabling of the process
|
|
0:07:30
|
So, whereas we say, IPv6 Router
|
|
0:07:36
|
to enable the EIGRPv6 process.
|
|
0:07:39
|
Because the network statement can be confusing
|
|
0:07:43
|
It just controls what links
|
|
0:07:51
|
So, once we enable the protocol,
|
|
0:07:53
|
if we look at the Show IP EIGRP Interfaces,
|
|
0:07:58
|
or Show IP Protocols,
|
|
0:08:00
|
that's gonna tell us what particular local
|
|
0:08:05
|
If we look at Show IP
|
|
0:08:07
|
it's gonna show things
|
|
0:08:11
|
What are the adjacencies that we have
|
|
0:08:14
|
and other attributes on a per-link basis
|
|
0:08:19
|
Now, for some reason, we find that there's
|
|
0:08:24
|
That we cannot get a basic
|
|
0:08:27
|
The output we wanna look
|
|
0:08:31
|
Not any output that says Debug IP EIGRP.
|
|
0:08:36
|
The reason why is that although
|
|
0:08:41
|
it was originally written to route
|
|
0:08:46
|
It was able to route IPv4,
|
|
0:08:51
|
So, the Debug EIGRP is referring
|
|
0:08:57
|
where Debug IP EIGRP would be
|
|
0:09:02
|
But the establishment of the
|
|
0:09:06
|
It would be shown as the
|
|
0:09:09
|
Debug EIGRP Packet Acknowledgement,
|
|
0:09:12
|
that's gonna show us what happens
|
|
0:09:17
|
This would also tell us...
|
|
0:09:19
|
if there were any mismatches
|
|
0:09:22
|
like the authentication or the metric waiting
|
|
0:09:28
|
Debug EIGRP Packet is gonna
|
|
0:09:35
|
Next thing we would verify is the
|
|
0:09:39
|
If we look at the Show IP EIRGP Neighbors
|
|
0:09:43
|
the key output here we're looking for
|
|
0:09:46
|
is that the update queue count is zero.
|
|
0:09:51
|
The queue count indicates how many
|
|
0:09:55
|
in order to be sent to
|
|
0:09:59
|
So, in other words, if we see this
|
|
0:10:02
|
it means that the routers are currently
|
|
0:10:07
|
And if the number continues
|
|
0:10:10
|
it means that there's something wrong
|
|
0:10:14
|
Normally, the actual update process,
|
|
0:10:19
|
should be very, very fast.
|
|
0:10:22
|
If we see that it is a non-zero value,
|
|
0:10:26
|
we could potentially have
|
|
0:10:28
|
or maybe one router is able to
|
|
0:10:33
|
but the other neighbor is not.
|
|
0:10:35
|
Like maybe over a frame-relay
|
|
0:10:38
|
if we don't have the proper
|
|
0:10:41
|
or maybe some sort of Layer 2 Ethernet where
|
|
0:10:47
|
Like a VLAN access list or an access
|
|
0:10:51
|
that might be potentially dropping
|
|
0:10:58
|
Once we verify that the
|
|
0:11:01
|
we can look at the actual routing
|
|
0:11:05
|
by looking at the Show IP EIGRP
|
|
0:11:13
|
Now, in EIGRP, we'll see...
|
|
0:11:16
|
that similar to BGP,
|
|
0:11:18
|
we only advertise the pass
|
|
0:11:24
|
So, unlike OSPF that advertises
|
|
0:11:29
|
EIGRP will only advertise the ones that are
|
|
0:11:35
|
This means that in some designs,
|
|
0:11:37
|
if there is a reason that a route is in
|
|
0:11:43
|
it means that, that particular
|
|
0:11:47
|
And that is also the same case for RIP.
|
|
0:11:50
|
We'll see that when we get into redistribution
|
|
0:11:55
|
For EIGRP and RIP, if the route
|
|
0:12:00
|
it means that we cannot advertise it.
|
|
0:12:03
|
We'll see this when we look
|
|
0:12:06
|
If it says that "The feasible distance
|
|
0:12:12
|
It means that for some reason, that is not
|
|
0:12:16
|
which then ultimately means
|
|
0:12:26
|
So, in our topology, our first step...
|
|
0:12:29
|
is that we're going to enable the process.
|
|
0:12:31
|
And we'll do it just on every interface
|
|
0:12:35
|
with the same autonomous system
|
|
0:12:39
|
So, in a similar configuration to RIP,
|
|
0:12:41
|
I'm simply gonna send all of the router EIGRP
|
|
0:12:49
|
So, on the access server,
|
|
0:12:53
|
Send Star (*) to go on all the lines.
|
|
0:12:56
|
We'll tell the routers to go to global config.
|
|
0:12:59
|
I wanna make sure that all the
|
|
0:13:03
|
We'll run the EIGRP process.
|
|
0:13:07
|
And then...
|
|
0:13:09
|
specify that any interface that starts with 155.10
|
|
0:13:15
|
we'll run EIGRP along with the loopback
|
|
0:13:26
|
Now, I could have also said Network 0.0.0.0
|
|
0:13:34
|
that would mean that any interfaces
|
|
0:13:38
|
that the EIGRP process would be enabled.
|
|
0:13:42
|
So, it really just depends how granular do
|
|
0:13:47
|
In a production network, you want
|
|
0:13:49
|
So, you don't want someone changing
|
|
0:13:55
|
And then, enabling them into the process unless
|
|
0:14:01
|
Now, within the scope of the lab exam, it really
|
|
0:14:05
|
So, if they don't care how you
|
|
0:14:09
|
The network statement that is exact
|
|
0:14:12
|
Or one that is vague that
|
|
0:14:16
|
The only problem with
|
|
0:14:20
|
Network 0.0.0.0 255.255.255.255 is that if I
|
|
0:14:31
|
let's say that a loopback for
|
|
0:14:35
|
it means that it's automatically gonna
|
|
0:14:41
|
And based on the individual design in the exam,
|
|
0:14:47
|
So, you technically can do this. I would
|
|
0:14:50
|
I would do the more specific
|
|
0:14:55
|
either that are based on the mask of
|
|
0:15:02
|
Now, in any case, as long as we look
|
|
0:15:07
|
and this shows the correct links
|
|
0:15:10
|
that's really the ultimate goal.
|
|
0:15:14
|
So, regardless of the syntax
|
|
0:15:16
|
this is really what we care about. So, is EIGRP
|
|
0:15:22
|
We can also see from this output.
|
|
0:15:25
|
It says, "The number of peers that
|
|
0:15:31
|
So, if the interface is a transit link
|
|
0:15:35
|
that it says that there are zero peers,
|
|
0:15:37
|
it would mean there's something wrong
|
|
0:15:42
|
On router 1's Fast Ethernet 0/0,
|
|
0:15:45
|
it says that there are two peers,
|
|
0:15:48
|
which in this case are the adjacency
|
|
0:15:56
|
and from 1 to 6.
|
|
0:15:58
|
So, this is what we would
|
|
0:16:01
|
Whereas with the serial 0/0.1, that's the
|
|
0:16:07
|
That's talking about router 5.
|
|
0:16:10
|
Then, the other point-to-point link going
|
|
0:16:15
|
So, simply based on the
|
|
0:16:19
|
it's already giving me a good amount of information
|
|
0:16:23
|
So, the adjacencies between router 1 and these
|
|
0:16:32
|
Now, if we look at the Show
|
|
0:16:36
|
it's gonna give me some more
|
|
0:16:40
|
state of the interface like,
|
|
0:16:45
|
It says that the authentication mode is not set.
|
|
0:16:48
|
So, we're not running MD-5 authentication.
|
|
0:16:51
|
It says that we can use multicast on this link.
|
|
0:16:57
|
So, depending on the underlying media,
|
|
0:17:01
|
it's gonna choose whether we want
|
|
0:17:05
|
For the LAN interface, we should be using multicast,
|
|
0:17:11
|
We see, for the point-to-point serial link,
|
|
0:17:14
|
because we're assuming that there's
|
|
0:17:20
|
Now, there is no EIRGP network
|
|
0:17:26
|
The process itself should automatically
|
|
0:17:31
|
But the key point being is that by default, we need
|
|
0:17:36
|
in order for the adjacencies to establish.
|
|
0:17:40
|
If we filtered out the unicast, we would
|
|
0:17:44
|
but the topology does
|
|
0:17:47
|
So, we would not be able to install
|
|
0:17:56
|
Next thing, we would wanna look
|
|
0:18:01
|
So, from this output, it tells us
|
|
0:18:05
|
what are the interfaces
|
|
0:18:09
|
What's the hold time that is left whether
|
|
0:18:15
|
Then, most importantly,
|
|
0:18:20
|
So here, if the queue count is zero,
|
|
0:18:23
|
then, it means that these
|
|
0:18:28
|
Usually, the only case that you
|
|
0:18:32
|
Is that you coincidentally looked at
|
|
0:18:36
|
while they're actually doing
|
|
0:18:39
|
or more likely that there's some problem
|
|
0:18:44
|
Like if the unicasts are filtered out or
|
|
0:18:49
|
or maybe there's a K value mismatch,
|
|
0:18:52
|
anything where the queue count
|
|
0:18:56
|
means that the routers are not
|
|
0:19:07
|
Next, if we were to look at
|
|
0:19:12
|
these are the different packet
|
|
0:19:16
|
For the adjacency, we may wanna look at
|
|
0:19:21
|
If I just say, Debug EIGRP Packets,
|
|
0:19:26
|
So, we see that we're sending hellos,
|
|
0:19:31
|
This is a good basic verification to make sure that
|
|
0:19:38
|
So, if I'm sending hellos out the link,
|
|
0:19:43
|
then that potentially could be a problem.
|
|
0:19:49
|
Once the adjacencies are established,
|
|
0:19:51
|
then, we should exchange
|
|
0:19:54
|
and run the dual calculation to actually figure
|
|
0:19:59
|
If we look at the Show IP
|
|
0:20:04
|
this is gonna show us all
|
|
0:20:07
|
Or not the currrent links, the current routes
|
|
0:20:14
|
Now, it doesn't necessarily mean that all
|
|
0:20:20
|
because normally, only the prefix that
|
|
0:20:26
|
which essentially is the
|
|
0:20:30
|
is gonna get installed into the routing table.
|
|
0:20:34
|
So, from router 1's perspective,
|
|
0:20:38
|
the link between router 3 and switch 1,
|
|
0:20:43
|
it says, "I have two possible routes to get there.
|
|
0:20:48
|
If I go to router 6, my
|
|
0:20:52
|
If I go to router 3, my total
|
|
0:20:57
|
So, I'm obviousluy going to prefer
|
|
0:21:02
|
So, that means that's what
|
|
0:21:06
|
So, the calculation itself is
|
|
0:21:10
|
You simply just look at what are all the
|
|
0:21:14
|
and pick the one that
|
|
0:21:17
|
Okay, it's only when we really
|
|
0:21:19
|
doing the backup paths with
|
|
0:21:22
|
Then, trying to do any path
|
|
0:21:27
|
values that make up the composite metric,
|
|
0:21:30
|
then, it starts to get kind of complicated
|
|
0:21:34
|
But under most circumstances,
|
|
0:21:38
|
and EIGRP is gonna take care of itself.
|
|
0:21:45
|
Now, the other output that we might wanna see
|
|
0:21:53
|
which will show us also the
|
|
0:21:59
|
Now, these are the ones that could
|
|
0:22:06
|
that we are normally
|
|
0:22:10
|
because we cannot guarantee
|
|
0:22:13
|
that the traffic is not
|
|
0:22:16
|
So, if we're to look at that...
|
|
0:22:20
|
37 route for example.
|
|
0:22:24
|
37 actually only has two paths still.
|
|
0:22:31
|
previously had only...
|
|
0:22:37
|
a few paths. Let's see, the 23.
|
|
0:22:47
|
Okay, the 23, when we look at the Show
|
|
0:22:51
|
155.10.23.0/24.
|
|
0:22:58
|
Notice that this says that
|
|
0:23:02
|
Okay, the total metric for
|
|
0:23:07
|
If we look at the Show IP
|
|
0:23:16
|
it says, "In reality, there's two
|
|
0:23:21
|
but one of them is not valid, because we
|
|
0:23:28
|
Now, the way that we do this
|
|
0:23:34
|
that the values EIGRP uses for it
|
|
0:23:40
|
Okay, the basic logic is that
|
|
0:23:45
|
what metric are you advertising
|
|
0:23:50
|
and what is my metric to reach you?
|
|
0:23:52
|
I'm gonna take what you're advertising
|
|
0:23:57
|
Add those two together, that's
|
|
0:24:02
|
If I see some other neighbor that is
|
|
0:24:07
|
that the value they're advertising to me
|
|
0:24:16
|
It means that they have to be a loop-free path,
|
|
0:24:23
|
So if we look at this visually,
|
|
0:24:26
|
let's say for example that router 1
|
|
0:24:28
|
has two possible neighbors
|
|
0:24:34
|
Now, beyond this, router 1 has no visibility of
|
|
0:24:40
|
because in EIRGP, it doesn't
|
|
0:24:43
|
It only knows about what the directly
|
|
0:24:47
|
So, router 2 says...
|
|
0:24:50
|
that my metric to get
|
|
0:24:55
|
Now, maybe router 1's metric
|
|
0:25:01
|
So, the total cost in this path is 30.
|
|
0:25:05
|
Router 3 is advertising a cost of 5,
|
|
0:25:10
|
but my link to get to router 3 is a
|
|
0:25:16
|
Now, we obviously know
|
|
0:25:19
|
so we would try to prefer
|
|
0:25:22
|
that's gonna have
|
|
0:25:26
|
But now, when we compare the value
|
|
0:25:31
|
versus what router 3 is advertising.
|
|
0:25:38
|
Since our metric is greater than
|
|
0:25:43
|
there is no possible way that
|
|
0:25:50
|
Because if it were, then, their
|
|
0:25:56
|
So, based on the simple logic that EIGRP says that
|
|
0:26:03
|
you're actually closer to
|
|
0:26:06
|
And this is what we call the
|
|
0:26:10
|
So, we'll see when we actually get
|
|
0:26:14
|
this is how EIGRP can get
|
|
0:26:17
|
by having pre-calculated loop-free
|
|
0:26:24
|
So, in this potential case, if router 1
|
|
0:26:29
|
it would immediately start
|
|
0:26:32
|
without having to do
|
|
0:26:35
|
And the reason why is that it knows that router 3
|
|
0:26:48
|
Now, when we look at this
|
|
0:26:51
|
the values that we would wanna look at...
|
|
0:26:54
|
Are the first number in the
|
|
0:27:01
|
Now, the first one is the total metric that
|
|
0:27:08
|
So, it is whatever metric that they are
|
|
0:27:14
|
What the router does first is
|
|
0:27:18
|
The first one in the parentheses
|
|
0:27:22
|
Now, in this case, we see that
|
|
0:27:26
|
is lower than the 21 million that's
|
|
0:27:32
|
So, it means that this top value wins.
|
|
0:27:36
|
So feasible distance, that's simply the metric.
|
|
0:27:39
|
The next thing we do is look at what are
|
|
0:27:47
|
That's what the second value is.
|
|
0:27:49
|
So, router 3 is telling router 1 that its metric
|
|
0:28:01
|
Now, since 20 million is still higher
|
|
0:28:07
|
I cannot guarantee that router 3 is
|
|
0:28:13
|
So, this second path is then
|
|
0:28:17
|
which is why when we look at the
|
|
0:28:21
|
only the first value is showing up.
|
|
0:28:28
|
So again, if you wanna see what are all
|
|
0:28:31
|
not only the ones that
|
|
0:28:35
|
that's what Show IP EIGRP Topology
|
|
0:28:43
|
Then, if we wanna see what are the specifics
|
|
0:28:48
|
If we were to say Show IP EIGRP
|
|
0:28:57
|
this shows us all the possible
|
|
0:29:01
|
Plus all of the individual attributes
|
|
0:29:05
|
So, things like the minimum
|
|
0:29:07
|
the delay, the total reliability, the total load,
|
|
0:29:15
|
Now, we see, not all of these values are
|
|
0:29:20
|
It depends on how the routers
|
|
0:29:25
|
that go into the actual calculation formula.
|
|
0:29:28
|
And actually, by default, only
|
|
0:29:32
|
which is the minimum
|
|
0:29:36
|
These other ones, the reliability,
|
|
0:29:38
|
these are not used as part of the metric calculation.
|
|
0:29:42
|
Now, we'll see, the reliability, and the load technically,
|
|
0:29:48
|
but it is not by default.
|
|
0:29:51
|
The MTU and the hop count, those are
|
|
0:29:56
|
MTU is used as a tie breaker,
|
|
0:29:57
|
where hop count is just to make
|
|
0:30:03
|
Because just like RIP, EIRGP technically
|
|
0:30:08
|
because it is a distance vector protocol.
|
|
0:30:12
|
information about what the
|