|
0:00:13
|
In our next section we're going to look at the other two first
|
|
0:00:16
|
hop redundancy protocols which are VRRP and GLBP
|
|
0:00:21
|
and essentially VRRP's functionality maps almost
|
|
0:00:24
|
identically one to one with HSRP. The main difference
|
|
0:00:29
|
being that this is an open standard based on RFC
|
|
0:00:32
|
5798 which we can see is for both IP version 4
|
|
0:00:36
|
and IP version 6
|
|
0:00:39
|
and we are using the terms master and backup
|
|
0:00:42
|
as opposed to primary and standby.
|
|
0:00:46
|
Now syntax wise, VRRP GLBP and HSRP are almost
|
|
0:00:50
|
identical where HSRP is using the standby command
|
|
0:00:54
|
VRRP is going to use the VRRP command
|
|
0:00:57
|
and GLBP is going to use the GLBP command.
|
|
0:01:01
|
We'll see that you can take the configuration of
|
|
0:01:04
|
one feature and map it almost identically to another
|
|
0:01:07
|
feature as long as you swap out the standby command
|
|
0:01:10
|
with VRRP or GLBP.
|
|
0:01:14
|
Some of the other key differences are that
|
|
0:01:16
|
VRRP is preemptive by default where HSRP is
|
|
0:01:20
|
not, so this means that if we are doing any type of
|
|
0:01:24
|
interface or enhanced object tracking and we want to choose
|
|
0:01:28
|
which is the master router as opposed to the active
|
|
0:01:31
|
based on the priority, then whoever has the
|
|
0:01:34
|
higher priority value is automatically going to
|
|
0:01:36
|
be elected. With HSRP we need to manually enable
|
|
0:01:40
|
preemption on whatever device we want to take over
|
|
0:01:44
|
the active status. Additionally the virtual
|
|
0:01:47
|
Mac address is going to use a different format, but just like
|
|
0:01:50
|
HSRP, the group number is going to be encoded
|
|
0:01:54
|
in the Mac address.
|
|
0:01:56
|
So for load balancing of the VRRP, we would have
|
|
0:01:59
|
multiple groups on the interface where some hosts are defaulting
|
|
0:02:03
|
to the virtual address of one group where other hosts
|
|
0:02:07
|
are defaulting to the virtual address of other groups.
|
|
0:02:10
|
So let's take a look at the command line configuration
|
|
0:02:13
|
here where we're going to take essentially identically what we had
|
|
0:02:17
|
on Router 4 and 6
|
|
0:02:19
|
with HSRP and we'll configure this instead
|
|
0:02:23
|
with VRRP.
|
|
0:02:25
|
So to start on Router 4, I'm going to remove our current
|
|
0:02:28
|
configuration. On Fast Ethernet 0/1 I'll say no standby 0
|
|
0:02:33
|
and no standby 1
|
|
0:02:38
|
which should remove all of my options
|
|
0:02:41
|
which it has and we're going to replace this with the
|
|
0:02:43
|
VRRP configuration. On Router 4 we'll say now instead
|
|
0:02:48
|
vrrp 1
|
|
0:02:53
|
The IP is 155.28.146.254
|
|
0:03:03
|
The VRR 1 -- vrrp 1 priority
|
|
0:03:09
|
will be 254
|
|
0:03:13
|
and we'll leave everything else as the default for now.
|
|
0:03:16
|
So on Router 6, we'll do the same thing.
|
|
0:03:20
|
On Fast Ethernet 0/0.146
|
|
0:03:24
|
no standby 1
|
|
0:03:28
|
instead we'll say vrrp 1
|
|
0:03:32
|
We see Router 6 is changing to the backup status
|
|
0:03:35
|
because Router 4 has the higher priority value
|
|
0:03:38
|
that I configured
|
|
0:03:39
|
if we look at now the show vrrp
|
|
0:03:43
|
as opposed to the show standby, Router 6 says that
|
|
0:03:46
|
it is the backup router.
|
|
0:03:50
|
This is the virtual address we've configured.
|
|
0:03:53
|
The virtual Mac address is 0000.5e00.0101
|
|
0:03:59
|
so again, where the last byte is encoding the group identifier.
|
|
0:04:03
|
Preemption is on by default, priority is still 100 by default
|
|
0:04:11
|
The master advertisement interval is one second
|
|
0:04:14
|
where the down interval is three seconds, so the default
|
|
0:04:17
|
timers are a little bit faster here, but again typically you would want
|
|
0:04:21
|
to rely on BFD, your Bidirectional Forwarding Detection
|
|
0:04:26
|
and not your Layer 3 keepalives
|
|
0:04:30
|
because it's generally going to take more CPU cycles
|
|
0:04:33
|
in the control plane to do keepalives that are for
|
|
0:04:36
|
VRRP or OSPF or BGP
|
|
0:04:39
|
those type of higher layer protocols as opposed to just
|
|
0:04:43
|
doing the light weight Layer 2 check
|
|
0:04:45
|
which is what BFD is designed to do.
|
|
0:04:50
|
If we look at the end result of this, from the end host it's
|
|
0:04:52
|
going to be completely transparent from them
|
|
0:04:55
|
where as we trace to 112.0.0.1
|
|
0:04:57
|
we can see the traffic is forwarding out to Router 4
|
|
0:05:02
|
If Router 4's link directly goes down
|
|
0:05:08
|
so this is the interface that we're actually running VRRP on
|
|
0:05:11
|
we should see that Router 6 promotes itself to the master
|
|
0:05:14
|
status and now the traffic is transparently redirected to them.
|
|
0:05:21
|
On Router 1, if we look at the show arp
|
|
0:05:26
|
and the virtual address is still going to point to that same
|
|
0:05:29
|
virtual Mac address regardless of who is servicing it.
|
|
0:05:32
|
The key being is that the master router just like the active router
|
|
0:05:36
|
in HSRP is the one that is it is going to send the
|
|
0:05:39
|
arp replies for any request and the gratuitous arps
|
|
0:05:44
|
which is the unsolicited reply to update both the CAM table
|
|
0:05:47
|
and the arp cache of the devices on the segment.
|
|
0:05:50
|
Now on Router 6 if we look at the debug IP packet
|
|
0:05:54
|
detail using access list 101
|
|
0:05:57
|
where again, access list 101 is saying not eigrp and not bgp
|
|
0:06:03
|
The difference between HSRP and VRRP from a transport
|
|
0:06:07
|
point of view is that we have a dedicated IP protocol
|
|
0:06:11
|
number which is 112 and we have a dedicated
|
|
0:06:16
|
multicast address which is 224.0.0.18
|
|
0:06:19
|
so beyond the differences from the packet level
|
|
0:06:23
|
format where obviously we can't run HSRP on one
|
|
0:06:26
|
side and VRRP on the other side because they're
|
|
0:06:28
|
two different protocols. The functional difference is going to be
|
|
0:06:31
|
pretty much the same.
|
|
0:06:34
|
Where from a design point of view, you would probably want to
|
|
0:06:36
|
choose VRRP over GLBP
|
|
0:06:39
|
just in case in the future you need to do any sort of
|
|
0:06:42
|
vendor interoperability since VRRP is the open standard
|
|
0:06:46
|
then generally most of the time it's going to be better to
|
|
0:06:49
|
go towards the standard's protocols as opposed to the
|
|
0:06:53
|
the vendor proprietary ones.
|
|
0:06:57
|
Now when we look at the next iteration which is through GLBP
|
|
0:07:02
|
The advantage of using this one is that we can do
|
|
0:07:06
|
multiple active routers at the same time
|
|
0:07:11
|
in order to do load balancing out of the network.
|
|
0:07:14
|
Where GLBP is a Cisco proprietary protocol
|
|
0:07:17
|
again just like HSRP, but it extends the HSRP
|
|
0:07:21
|
functionality by decoupling the role of the person in charge
|
|
0:07:26
|
of the ARP responses
|
|
0:07:30
|
versus the person that actually forwards the traffic
|
|
0:07:32
|
based on the virtual Mac addresses.
|
|
0:07:36
|
Where in both HSRP and VRRP, we only have the
|
|
0:07:39
|
single roles of the active router or the master router
|
|
0:07:43
|
GLBP has two separate roles where these are called the
|
|
0:07:46
|
active virtual gateways and the active virtual forwarders.
|
|
0:07:52
|
The active virtual forwarders are the routers that have
|
|
0:07:54
|
the virtual Mac addresses assigned that actually
|
|
0:07:58
|
forward the packets out of the network.
|
|
0:08:01
|
The key is that we have this additional role
|
|
0:08:03
|
which is now the active virtual gateway
|
|
0:08:06
|
that is in charge of responding to the ARP requests from the
|
|
0:08:09
|
end hosts.
|
|
0:08:12
|
And the key point is that the active virtual gateway
|
|
0:08:15
|
is going to respond with the Mac address of the active
|
|
0:08:17
|
virtual forwarder.
|
|
0:08:20
|
So depending on how we have the load balancing
|
|
0:08:22
|
configured whether it's on a one to one basis
|
|
0:08:25
|
or whether it's based on a specific weighting value
|
|
0:08:28
|
that is going to control how many times the active virtual
|
|
0:08:32
|
gateway uses the Mac address of one host versus the other
|
|
0:08:36
|
in the ARP reply.
|
|
0:08:41
|
So just like other load balancing mechanisms, it's hard to get an
|
|
0:08:44
|
equal distribution where you would never have a traffic
|
|
0:08:47
|
flow that's exactly one to one or exactly two to one
|
|
0:08:51
|
with your weighting because we're not doing a packet level
|
|
0:08:54
|
load balancing or even a destination load balancing
|
|
0:08:58
|
we're doing this based on the ARP replies.
|
|
0:09:02
|
So depending on the underlying traffic pattern of the end hosts
|
|
0:09:07
|
we could get somewhere close to the weighting that we're
|
|
0:09:10
|
configuring, but it's never going to be an exact value.
|
|
0:09:13
|
So typically when you actually implement this, you would want
|
|
0:09:15
|
to do some sort of application monitoring like NetFlow or
|
|
0:09:20
|
SNMP monitoring to figure out what's the actual
|
|
0:09:24
|
result of the utilization based on load balancing, then you could
|
|
0:09:27
|
change the weighting around in order to get whatever
|
|
0:09:30
|
particular value that you want.
|
|
0:09:33
|
Now implementation wise, the key difference again is that we
|
|
0:09:37
|
have this decoupling of the active virtual gateway
|
|
0:09:40
|
from the active virtual forward.
|
|
0:09:42
|
The active virtual gateway is the one that is replying
|
|
0:09:46
|
to the ARP requests.
|
|
0:09:48
|
So there's going to be one gateway where all of the other
|
|
0:09:52
|
routers are going to be backing up the virtual gateway.
|
|
0:09:57
|
The active virtual gateway by default is going to be
|
|
0:09:59
|
the only active virtual forwarder unless we
|
|
0:10:03
|
configure the other routers to do load balancing.
|
|
0:10:07
|
So now let's take our next configuration here which was
|
|
0:10:10
|
the VRRP and we'll change this to match the GLBP config.
|
|
0:10:17
|
So on Router 4 let's show run interface Fast Ethernet 0/1
|
|
0:10:22
|
We'll say no vrrp 1
|
|
0:10:26
|
and we'll replace this with the glbp 1
|
|
0:10:31
|
The IP address is going to be the same 155.28.146.254
|
|
0:10:42
|
on Router 6 likewise
|
|
0:10:45
|
we'll say no vrrp 1
|
|
0:10:49
|
and set that to GLBP
|
|
0:10:54
|
If we now look at the show GLBP
|
|
0:10:57
|
as opposed to show standby or show HSRP
|
|
0:11:03
|
We should see once the routers go through their initial
|
|
0:11:06
|
convergence, we discover who is the active virtual gateway
|
|
0:11:11
|
and who are the active virtual forwarders.
|
|
0:11:17
|
Router 6 says that preemption is disabled.
|
|
0:11:24
|
The active router is Router 4
|
|
0:11:27
|
has a priority of 100
|
|
0:11:29
|
This is talking about the active virtual gateway.
|
|
0:11:34
|
So when we keep going a little bit further in this output
|
|
0:11:37
|
it says that there are two forwarders which consist
|
|
0:11:43
|
of this particular Mac address
|
|
0:11:47
|
which is Router 4
|
|
0:11:50
|
and the second Mac address F922
|
|
0:11:54
|
which is us locally is Router 6
|
|
0:11:59
|
then it says we changed to the listen state which means that
|
|
0:12:01
|
Router 4 is now active.
|
|
0:12:03
|
If we show glbp 1
|
|
0:12:08
|
or just show glbp
|
|
0:12:13
|
where again, this active status this is talking about the
|
|
0:12:16
|
active virtual gateway.
|
|
0:12:19
|
So technically, it doesn't even really matter who is the active
|
|
0:12:22
|
virtual gateway
|
|
0:12:23
|
as long as the active virtual forwarders are correct
|
|
0:12:26
|
which is the devices that are actually forwarding the
|
|
0:12:30
|
traffic onto the link.
|
|
0:12:32
|
Right now it says that forwarder 1 is active
|
|
0:12:35
|
where forwarder 2 is listen.
|
|
0:12:38
|
Again, we need then to do the correlation between
|
|
0:12:40
|
the Mac addresses and the actual IPs.
|
|
0:12:43
|
So the 9B address this is Router 4
|
|
0:12:48
|
that is the first forwarder.
|
|
0:12:50
|
This is active
|
|
0:12:54
|
where the second forwarder 22 that's Router 6's address
|
|
0:12:58
|
right now this is in the listen state.
|
|
0:13:02
|
So if we look at the result of this and go to switch 4
|
|
0:13:06
|
and do a trace route out to that same destination
|
|
0:13:10
|
we should see ultimately that the traffic is going to be
|
|
0:13:12
|
received by Router 4
|
|
0:13:15
|
Now what I may need to do here on the end host that's
|
|
0:13:19
|
doing the decision which essentially is Router 1
|
|
0:13:22
|
here when I look at the show arp, I need to make sure
|
|
0:13:26
|
that that old Mac address is flushed out of the
|
|
0:13:30
|
CAM table and out of the ARP cache.
|
|
0:13:34
|
So some of the potential convergence problems with
|
|
0:13:36
|
these first hop redundancy protocols are going to be the IP address to
|
|
0:13:40
|
Mac address mapping and then also the Mac address
|
|
0:13:43
|
in the CAM table.
|
|
0:13:45
|
So right now, traffic is being dropped basically
|
|
0:13:49
|
because Router 1 has the wrong ARP information.
|
|
0:13:53
|
If I tell it to clear out the ARP cache
|
|
0:13:57
|
I should see the .254 address is going to change
|
|
0:14:02
|
to now the Mac address that is specific to GLBP
|
|
0:14:07
|
It says 0007.b400.0101
|
|
0:14:10
|
Specifically if we correlate this with what Router 4 says
|
|
0:14:17
|
this 0101 is the Mac address of Router 4
|
|
0:14:22
|
as the active virtual forwarder.
|
|
0:14:28
|
So we really don't care about what is the Mac
|
|
0:14:30
|
address of the active virtual gateway because that
|
|
0:14:33
|
is not the device that's actually forwarding the packets.
|
|
0:14:35
|
The active virtual gateway is basically used just for
|
|
0:14:38
|
the control plane of this feature.
|
|
0:14:42
|
Now if we send packets out to the destination
|
|
0:14:46
|
it says that they're going out to Router 4
|
|
0:14:49
|
Now to test the actual load balancing here
|
|
0:14:52
|
what I'm going to do on Router 4 and Router 6
|
|
0:14:54
|
is on their inbound connections to this
|
|
0:14:58
|
VLAN 146, I'm going to configure an access list
|
|
0:15:01
|
that is going to account for our ICMP echoes
|
|
0:15:07
|
which means we're going to be using just basic pings
|
|
0:15:10
|
in order to test which of the exit points are we using.
|
|
0:15:14
|
So on Router 4 I'm going to configure a new access list
|
|
0:15:18
|
we'll say access list 111 permits icmp any any
|
|
0:15:23
|
access list 111 is then going to permit ip any any
|
|
0:15:28
|
and this is going to apply inbound on Fast Ethernet 0/1
|
|
0:15:34
|
ip access group 111 in
|
|
0:15:36
|
Router 6 will essentially do the same thing.
|
|
0:15:41
|
Then we can use the show access list output
|
|
0:15:45
|
to figure out which gateway is actually being used for
|
|
0:15:49
|
the forwarding.
|
|
0:15:51
|
ip access group 111 in
|
|
0:15:57
|
So now from the end hosts let's say on all the switches
|
|
0:16:01
|
switch 1 through switch 4, we'll do pings out to these
|
|
0:16:04
|
destinations, we'll say 112.0.0.1
|
|
0:16:07
|
and we'll send 10 packets.
|
|
0:16:10
|
We'll do this on each of the four switches.
|
|
0:16:15
|
So now when I look at my packet counters
|
|
0:16:19
|
there have been 40 packets total. On Router 4 let's look
|
|
0:16:23
|
at the show access list
|
|
0:16:28
|
Router 4 says that 40 packets came in.
|
|
0:16:30
|
So this essentially means that with the default configuration
|
|
0:16:35
|
we're not doing any load balancing.
|
|
0:16:39
|
So we do need to go the step further to configure
|
|
0:16:42
|
the active virtual gateway to tell it what is the weighting going to be
|
|
0:16:46
|
which is essentially how it responds to the ARP replies.
|
|
0:16:55
|
Again, specifically this is Router 4 who is elected the gateway.
|
|
0:16:59
|
So we'll say on this link for GLBP 1
|
|
0:17:03
|
I want to change the weighting
|
|
0:17:07
|
Let's say that my weighting value is 10
|
|
0:17:14
|
then we can change the different threshold values
|
|
0:17:19
|
this is going to determine whether it's actually being
|
|
0:17:21
|
used or not based on tracking.
|
|
0:17:23
|
So we can get into a more complex procedure with
|
|
0:17:26
|
tracking that if I have multiple outbound links
|
|
0:17:29
|
for my uplinks
|
|
0:17:31
|
let's say I have two GIG Ethernet links.
|
|
0:17:34
|
I could say that if both of them are up, I'm going to
|
|
0:17:37
|
have a weighting of 10, but if one of those -- one of them
|
|
0:17:40
|
goes down, maybe I'll reduce to a weighting of five.
|
|
0:17:43
|
If both of them are down I want to remove myself
|
|
0:17:47
|
from the active virtual forwarder election
|
|
0:17:49
|
because I don't have anywhere to get out to the rest of the network.
|
|
0:17:53
|
So now I have the weighting configured, I'm also going to
|
|
0:17:56
|
say that the specific load balancing is going to use
|
|
0:18:00
|
the weighting.
|
|
0:18:02
|
So load balance and proportion to the forwarder's
|
|
0:18:04
|
weighting where Router 4 is configured as ten
|
|
0:18:07
|
I'm going to configure Router 6 then as a weighting of five.
|
|
0:18:14
|
GLBP 1 the weighting is five.
|
|
0:18:21
|
So now on Router 4 if we look at the show glbp
|
|
0:18:25
|
it says that there are two forwarders still only one of
|
|
0:18:28
|
them is active.
|
|
0:18:32
|
So we need to see ultimately that Router 6 is going to become
|
|
0:18:36
|
an active forwarder
|
|
0:18:38
|
which typically what you would do with this configuration
|
|
0:18:42
|
is just make sure that they all have the same config.
|
|
0:18:46
|
So Router 6 would agree that they're going to do the load balancing
|
|
0:18:50
|
based on the weights
|
|
0:18:56
|
then if we show glbp
|
|
0:19:04
|
it says for forwarder 2 we are active.
|
|
0:19:11
|
So now let's try this with a different destination let's
|
|
0:19:13
|
ping 113.0.0.1
|
|
0:19:19
|
we'll do this on all four of the switches.
|
|
0:19:27
|
When we look at the result of the show access lists
|
|
0:19:36
|
Router 4 says I have a thousand matches
|
|
0:19:38
|
actually this counter's on let's say clear access list counters
|
|
0:19:48
|
on both of these clear access list counters
|
|
0:19:51
|
let's try these pings again.
|
|
0:19:59
|
Then we will show access list.
|
|
0:20:22
|
And what I actually should have said here for the
|
|
0:20:29
|
ACL I should have said icmp any any echo
|
|
0:20:32
|
because some of these matches are from other
|
|
0:20:34
|
icmp type codes that are forwarding on the link
|
|
0:20:39
|
which mainly are going to be redirects here.
|
|
0:20:42
|
So let's do this real quick let's say ip access list 111
|
|
0:20:50
|
or extended 111 no 10
|
|
0:20:54
|
10 says permit icmp any any echo
|
|
0:21:00
|
So we're replacing that line to be a little bit more specific.
|
|
0:21:04
|
Then I'll do the same thing on Router 6
|
|
0:21:06
|
ip access list extended 111
|
|
0:21:10
|
10 is changing just to echo
|
|
0:21:12
|
if we show access list
|
|
0:21:16
|
on both of them
|
|
0:21:32
|
we sent packets to the destination if we now look at the show access list
|
|
0:21:37
|
Router 4 received 25 of the packets
|
|
0:21:43
|
Router 6 didn't receive any
|
|
0:21:46
|
which means that Router 4's counter is just lagging here.
|
|
0:21:49
|
So this should essentially go up.
|
|
0:21:52
|
What this essentially means is that even though there's
|
|
0:21:56
|
multiple sources that are forwarding traffic onto the
|
|
0:21:58
|
link, keep in mind that the load balancing is based on the
|
|
0:22:02
|
ARP request and reply.
|
|
0:22:07
|
So with our particular design here who is doing the ARP request
|
|
0:22:11
|
for the virtual IP address.
|
|
0:22:16
|
It's not the actual originators of the traffic because I'm originating
|
|
0:22:19
|
the traffic from the
|
|
0:22:28
|
the different switches, so traffic is coming in here
|
|
0:22:30
|
it's coming also from switch 2
|
|
0:22:34
|
it's coming from switch 1
|
|
0:22:36
|
and it's coming from switch 3
|
|
0:22:39
|
When the traffic actually gets onto this VLAN 146 segment
|
|
0:22:44
|
the device that's controlling where it's going to forward to
|
|
0:22:47
|
is Router 1
|
|
0:22:49
|
because Router 1 has the ARP entry that goes
|
|
0:22:52
|
from the virtual address to the virtual Mac address.
|
|
0:22:59
|
So what I would need to do is actually clear
|
|
0:23:01
|
the ARP cache on Router 1
|
|
0:23:03
|
and look at what changes
|
|
0:23:08
|
if I change from one virtual gateway to another one.
|
|
0:23:13
|
So from Router 1 let's ping 114.0.0.1
|
|
0:23:19
|
clear the ARP cache and then do this again.
|
|
0:23:22
|
Eventually what I should see is that the Mac address
|
|
0:23:26
|
is changing from the .101 address
|
|
0:23:33
|
to the dot -- or actually the .0101
|
|
0:23:39
|
to .0102
|
|
0:23:47
|
One of these is going to represent Router 4
|
|
0:23:49
|
one of these is going to represent Router 6
|
|
0:23:53
|
So again, the load balancing is not going to be exact
|
|
0:23:55
|
regardless of what the weighting you configure is
|
|
0:23:59
|
because it's not a per destination load balancing
|
|
0:24:02
|
like CEF would be, it's a per ARP request
|
|
0:24:06
|
load balancing.
|
|
0:24:08
|
So this means if host one gets gateway one
|
|
0:24:13
|
and host two and three get gateway two
|
|
0:24:17
|
if host one is still sending the majority of the traffic flows
|
|
0:24:20
|
then the weighting is not going to be in a ratio of one to two.
|
|
0:24:25
|
So again, it's kind of a hack on the process, but it
|
|
0:24:29
|
it at least allows you to forward through both of them at the same time.
|
|
0:24:32
|
So if we compare the show access list output on Router 4
|
|
0:24:36
|
versus Router 6
|
|
0:24:38
|
we see at least now that Router 6 is getting some of the
|
|
0:24:40
|
hits, but in general, they would be in a ratio of more
|
|
0:24:46
|
traffic being sent to Router 4 versus Router 6
|
|
0:24:49
|
because we have the weighting set to be higher.
|
|
0:24:51
|
The vast majority of the other features are going to be
|
|
0:24:55
|
nearly identical between the protocols.
|
|
0:24:57
|
So if we say glbp 1
|
|
0:25:00
|
we can do authentication with both md5 and clear text.
|
|
0:25:05
|
If I wanted to change the timers, this is going to
|
|
0:25:09
|
affect the reconvergence directly on that segment.
|
|
0:25:14
|
If I want to change my
|
|
0:25:19
|
weighting let's say weighting we want to do tracking
|
|
0:25:23
|
then I'll track object number 10
|
|
0:25:25
|
If object number 10 is down, I'm going to decrement my
|
|
0:25:28
|
weighting value as opposed to my priority
|
|
0:25:33
|
because the priority is affecting the active virtual
|
|
0:25:36
|
gateway election, not the active virtual forwarder election.
|
|
0:25:42
|
If we wanted to change the particular forwarder options
|
|
0:25:45
|
we could say glbp 1 forwarder
|
|
0:25:50
|
then I want to take over the state for other lower
|
|
0:25:54
|
priority forwarders, so this would be the case where you
|
|
0:25:56
|
only want one virtual forwarder similar to HSRP or VRRP
|
|
0:26:02
|
and you don't want to do load balancing between
|
|
0:26:05
|
forwarders that have different weight values.
|
|
0:26:08
|
So with this third option GLBP, you can get very
|
|
0:26:10
|
granular with it, but when you look at the comparison
|
|
0:26:14
|
of the final result, so the affected differences is not
|
|
0:26:19
|
really that much. HSRP and VRRP are almost identical
|
|
0:26:23
|
where GLBP is adding this load balancing mechanism
|
|
0:26:26
|
by separating the active virtual gateway versus the
|
|
0:26:29
|
active virtual forwarder. Now within the scope of the
|
|
0:26:32
|
exam for these three first hop redundancy protocols
|
|
0:26:35
|
as long as you try this out a couple times, then you
|
|
0:26:38
|
know where this is located in the configuration guides
|
|
0:26:41
|
which again is going to be under the IP application services.
|
|
0:26:44
|
Then the first hop redundancy protocols.
|
|
0:26:47
|
If you look at either the configuration examples
|
|
0:26:51
|
or the command reference
|
|
0:26:53
|
like let's say here configuration example customizing GLBP
|
|
0:26:58
|
you should essentially be able to take this config
|
|
0:27:01
|
and essentially just change it around for whatever values
|
|
0:27:04
|
that are being asked in the question. So it's
|
|
0:27:06
|
more of an issue of a design choice
|
|
0:27:09
|
as to which of the three do you want to use.
|
|
0:27:11
|
If load balancing is needed, then you're going to have to use
|
|
0:27:14
|
GLBP. For VRRP and HSRP they're the same
|
|
0:27:18
|
with the difference behind the scenes being that
|
|
0:27:22
|
VRRP is the open standard and HSRP is Cisco proprietary.
|
