WEBVTT

0:00:02.820000 --> 0:00:07.940000
 Hello and welcome to this video titled,
 Configuring Cisco iOS as a DHCP

0:00:07.940000 --> 0:00:10.920000
 server or relay agent.

0:00:10.920000 --> 0:00:16.080000
 In this video, I'm going to quickly
 talk about DHCP broadcast handling

0:00:16.080000 --> 0:00:20.120000
 and what routers would normally
 do with DHCP packets.

0:00:20.120000 --> 0:00:24.460000
 We'll talk about something about DHCP
 relay agents and how you can configure

0:00:24.460000 --> 0:00:32.300000
 a Cisco iOS device like a router
 or switch as a DHCP server.

0:00:32.300000 --> 0:00:35.920000
 So let's talk about DHCP broadcast
 handling for just a moment.

0:00:35.920000 --> 0:00:39.200000
 So we know that the messages sent from
 the client and here we're specifically

0:00:39.200000 --> 0:00:43.460000
 focusing on IP version 4 DHCP.

0:00:43.460000 --> 0:00:48.220000
 We know that your typical client messages
 are the discover and the request.

0:00:48.220000 --> 0:00:51.880000
 And we know that both of those go out
 with a destination address at both

0:00:51.880000 --> 0:00:55.420000
 layer 2 and layer 3 as
 a broadcast address.

0:00:55.420000 --> 0:01:00.260000
 Well, that's all well and good if your
 DHCP server is capable of receiving

0:01:00.260000 --> 0:01:04.760000
 that broadcast. In other words, if they
 sit on the same broadcast domain

0:01:04.760000 --> 0:01:09.160000
 or subnet that you do,
 but what if they don't?

0:01:09.160000 --> 0:01:11.540000
 Well, broadcasts are not routable.


0:01:11.540000 --> 0:01:16.440000
 So if the DHCP server is only reachable
 via three or four hops away, you

0:01:16.440000 --> 0:01:23.280000
 have to go through some routers to get to
 it, routers do not forward broadcasts.

0:01:23.280000 --> 0:01:28.440000
 So when a router or layer 3 switch
 receives a DHCP broadcast, there's

0:01:28.440000 --> 0:01:31.000000
 one of three ways it could react.

0:01:31.000000 --> 0:01:34.160000
 The default behavior
 is to inspect it.

0:01:34.160000 --> 0:01:38.180000
 So whenever a broadcast comes in, a
 router switch will first inspect it

0:01:38.180000 --> 0:01:40.700000
 and say, okay, is this
 broadcast for me?

0:01:40.700000 --> 0:01:44.980000
 Maybe somebody is trying to arp for
 me or if I'm running RIP version 1,

0:01:44.980000 --> 0:01:47.280000
 maybe someone's sending a
 RIP update or something.

0:01:47.280000 --> 0:01:51.540000
 Well, in this case, the router switch
 would say, oh, well, when I look

0:01:51.540000 --> 0:01:56.300000
 at the IP header here, the
 IP header says it's UDP.

0:01:56.300000 --> 0:02:01.080000
 I'm not running any UDP based applications
 that do broadcast.

0:02:01.080000 --> 0:02:03.940000
 So the router switch at that point
 would drop the broadcast.

0:02:03.940000 --> 0:02:06.260000
 It would discard it
 saying not for me.

0:02:06.260000 --> 0:02:10.360000
 Now, alternatively, we can configure
 that router switch to do one of two

0:02:10.360000 --> 0:02:15.120000
 other things. We could configure it to
 take that DHCP broadcast and relay

0:02:15.120000 --> 0:02:20.640000
 it or convey it onto the actual
 DHCP server itself.

0:02:20.640000 --> 0:02:23.960000
 And we're going to talk in just
 a moment about how to do that.

0:02:23.960000 --> 0:02:29.840000
 Or we could actually configure the router
 switch itself as a DHCP server.

0:02:29.840000 --> 0:02:32.820000
 Now, that's fine if you've got a really
 small network where like maybe

0:02:32.820000 --> 0:02:37.340000
 everybody in your building is just
 in one VLAN, one broadcast domain.

0:02:37.340000 --> 0:02:40.840000
 Then that case, sure, you could have a
 router switch in that same broadcast

0:02:40.840000 --> 0:02:43.600000
 domain and configure
 it as a DHCP server.

0:02:43.600000 --> 0:02:47.560000
 But if you've got any more than about
 three or four broadcast domains,

0:02:47.560000 --> 0:02:50.220000
 that's not a really
 scalable situation.

0:02:50.220000 --> 0:02:56.300000
 So let's talk about the second
 option here, DHCP relay agents.

0:02:56.300000 --> 0:03:04.500000
 So the idea behind this feature is that
 let's just focus on routers primarily.

0:03:04.500000 --> 0:03:09.320000
 So when a router receives a DHCP discover
 or a DHCP request, which is

0:03:09.320000 --> 0:03:11.900000
 a broadcast, normally
 just drops it.

0:03:11.900000 --> 0:03:16.300000
 Now with the relay agent functionality,
 we can actually tell that router

0:03:16.300000 --> 0:03:22.920000
 to take that broadcast and encapsulate
 it or change it into a unicast

0:03:22.920000 --> 0:03:27.900000
 packet. And then unicast it
 directly to the DHCP server.

0:03:27.900000 --> 0:03:31.200000
 This is what is meant by
 the term relay agent.

0:03:31.200000 --> 0:03:35.400000
 So the relay agent, when he does this,
 he's going to create a new packet

0:03:35.400000 --> 0:03:40.520000
 header. So the new IP header will have
 a source IP address of the relay

0:03:40.520000 --> 0:03:46.140000
 agent himself and a destination IP
 address not of a broadcast, but of

0:03:46.140000 --> 0:03:47.980000
 the actual DHCP server.

0:03:47.980000 --> 0:03:50.640000
 So we can see this right here.

0:03:50.640000 --> 0:03:52.920000
 So right there, there's the message
 coming in from the client.

0:03:52.920000 --> 0:03:59.580000
 So here's our original DHCP discover
 or possibly our DHCP request.

0:03:59.580000 --> 0:04:02.740000
 By default, source address
 is all zeros.

0:04:02.740000 --> 0:04:04.280000
 Destination address is broadcast.

0:04:04.280000 --> 0:04:08.300000
 Well, if we just left it that way,
 it wouldn't get beyond that line.

0:04:08.300000 --> 0:04:09.880000
 The router would drop it.

0:04:09.880000 --> 0:04:14.860000
 But by configuring the router as a relay
 agent, he can take that and repackage

0:04:14.860000 --> 0:04:17.300000
 it as a unicast packet.

0:04:17.300000 --> 0:04:22.320000
 So now when that packet is leaving his
 interface, it'll have the source

0:04:22.320000 --> 0:04:26.880000
 IP address of whatever
 interface received it.

0:04:26.880000 --> 0:04:31.320000
 Destination IP address of
 the DHCP server itself.

0:04:31.320000 --> 0:04:33.060000
 Now you say, well,
 how does he know?

0:04:33.060000 --> 0:04:36.640000
 How does the router know what the
 IP address is of the DHCP server?

0:04:36.640000 --> 0:04:39.340000
 You have to configure it.

0:04:39.340000 --> 0:04:41.160000
 And it's just one command.

0:04:41.160000 --> 0:04:45.100000
 Basically, you have to ask yourself,
 okay, what interfaces on this router

0:04:45.100000 --> 0:04:51.040000
 could I expect to receive
 DHCP broadcasts like this?

0:04:51.040000 --> 0:04:53.960000
 Oh, well, I can expect them
 on gigabit zero slash zero.

0:04:53.960000 --> 0:04:59.220000
 Okay, so on that interface, you just
 go to it and type one command.

0:04:59.220000 --> 0:05:02.780000
 It's the IP helper dash
 address command.

0:05:02.780000 --> 0:05:08.040000
 And then you put in the actual
 IP address of the DHCP server.

0:05:08.040000 --> 0:05:13.440000
 Now, if you have more than one DHCP server,
 you could replicate this command

0:05:13.440000 --> 0:05:16.680000
 two or three or four times, however
 many times you need with different

0:05:16.680000 --> 0:05:21.520000
 IP addresses. And then he'll take that
 one discover packet in, replicate

0:05:21.520000 --> 0:05:26.340000
 it, you know, n number of times, and
 then unicast it out to the different

0:05:26.340000 --> 0:05:31.280000
 DHCP servers. Now, there's one thing
 in here I also want to talk about

0:05:31.280000 --> 0:05:33.960000
 which I kind of skipped
 over, which was this.

0:05:33.960000 --> 0:05:39.120000
 You see, this DHCP server, he's probably
 got several pools, like he might

0:05:39.120000 --> 0:05:42.780000
 have a pool called
 human resources.

0:05:42.780000 --> 0:05:45.180000
 And that's where he's going to allocate
 addresses for people in the human

0:05:45.180000 --> 0:05:46.520000
 resources department.

0:05:46.520000 --> 0:05:52.800000
 You might have another pool of
 addresses called marketing.

0:05:52.800000 --> 0:05:57.440000
 And another pool of something
 else, maybe payroll.

0:05:57.440000 --> 0:06:04.580000
 So when relay agents out there are forwarding
 these DHCP packets to the

0:06:04.580000 --> 0:06:11.180000
 server, the server has to know which pool
 do I pick from to pick an address

0:06:11.180000 --> 0:06:14.080000
 and send it back to
 the original client.

0:06:14.080000 --> 0:06:19.380000
 Well, the way they know that is because
 when the relay agent took this

0:06:19.380000 --> 0:06:25.160000
 broadcast and converted it into a unicast
 in the first place, inside the

0:06:25.160000 --> 0:06:31.500000
 DHCP packet, there's a field known
 as the gateway IP address field.

0:06:31.500000 --> 0:06:35.000000
 Now, in some documents or papers you
 read, that might also be called the

0:06:35.000000 --> 0:06:37.600000
 relay agent IP address.

0:06:37.600000 --> 0:06:40.780000
 So it just depends on
 the documentation.

0:06:40.780000 --> 0:06:49.500000
 Now, when a client is creating a discover
 packet and he sends it out,

0:06:49.500000 --> 0:06:51.980000
 it has not gone through
 a gateway yet.

0:06:51.980000 --> 0:06:53.800000
 It hasn't touched a gateway.

0:06:53.800000 --> 0:06:59.000000
 So the gateway IP address
 field would be all zeros.

0:06:59.000000 --> 0:07:02.800000
 This is the field that's changed.

0:07:02.800000 --> 0:07:07.400000
 So as router one here is my relay agent
 gets it, not only is he going

0:07:07.400000 --> 0:07:11.480000
 to put his own IP address as the source
 IP address in the IP packet header

0:07:11.480000 --> 0:07:21.080000
 itself, he'll put his IP address
 into that gateway address field.

0:07:21.080000 --> 0:07:26.360000
 So the DHCP server, he's not really looking
 at the source address to figure

0:07:26.360000 --> 0:07:29.160000
 out which pool to use.

0:07:29.160000 --> 0:07:32.640000
 He's looking at the gateway
 IP address field.

0:07:32.640000 --> 0:07:37.060000
 And so he'll say, oh, I've got a
 pool here named human resources.

0:07:37.060000 --> 0:07:40.960000
 It's got a various range of addresses
 and in that pool, I've configured

0:07:40.960000 --> 0:07:46.760000
 a default router or a default
 gateway of 1.1.1.1.1.1.

0:07:46.760000 --> 0:07:52.280000
 And lo and behold, that's the exact same
 IP address that's in the gateway

0:07:52.280000 --> 0:07:54.800000
 IP address field of
 this DHCP packet.

0:07:54.800000 --> 0:08:00.720000
 So I guess I should pull an available
 address from this pool.

0:08:00.720000 --> 0:08:03.960000
 And I just want to show you
 here on the next slide.

0:08:03.960000 --> 0:08:07.100000
 We're not going to go into the gory details
 of this, but this is the DHCP

0:08:07.100000 --> 0:08:09.060000
 IP packet format.

0:08:09.060000 --> 0:08:11.660000
 And the main thing I want to call your
 attention to is that field we were

0:08:11.660000 --> 0:08:12.840000
 just talking about.

0:08:12.840000 --> 0:08:17.480000
 Right there. That is our
 gateway IP address field.

0:08:17.480000 --> 0:08:23.880000
 Normally it's all zeros when the client
 is creating his discover or request.

0:08:23.880000 --> 0:08:28.220000
 But when it goes through a relay agent,
 the relay agent puts his own IP

0:08:28.220000 --> 0:08:30.900000
 address in that field.

0:08:30.900000 --> 0:08:37.880000
 So that's very, very common, configuring
 a router as a relay agent.

0:08:37.880000 --> 0:08:42.220000
 You can also do that on a switch.

0:08:42.220000 --> 0:08:49.600000
 So for example, if I had something
 like this, here's my switch.

0:08:49.600000 --> 0:08:54.980000
 Here's my various switch ports
 leading out to my hosts.

0:08:54.980000 --> 0:08:58.440000
 Let's put little H's here
 from my laptops and PCs.

0:08:58.440000 --> 0:09:05.800000
 And let's say that all of these ports
 here were in a VLAN, like VLAN 2.

0:09:05.800000 --> 0:09:11.880000
 And then here is another interface which
 maybe leads to a router, which

0:09:11.880000 --> 0:09:14.320000
 eventually leads off
 to my DHCP server.

0:09:14.320000 --> 0:09:16.780000
 Let's give him the exact
 same address as before.

0:09:16.780000 --> 0:09:28.540000
 7777. Okay. And between these two guys
 right here, this is maybe 1.1.1

0:09:28.540000 --> 0:09:31.860000
.1 and this is 1.1.1.2.

0:09:31.860000 --> 0:09:36.480000
 So on this particular interface, we've
 done the No Switch port command.

0:09:36.480000 --> 0:09:40.780000
 In other words, we have turned
 it into a routed interface.

0:09:40.780000 --> 0:09:45.540000
 And then what we have for VLAN 2 is we've
 got a switched virtual interface.

0:09:45.540000 --> 0:09:54.800000
 We have interface VLAN
 2, IP address 2.2.2.2.

0:09:54.800000 --> 0:10:00.460000
 Well, this is where we would
 put our IP helper address.

0:10:00.460000 --> 0:10:05.980000
 IP helper-address 7777.

0:10:05.980000 --> 0:10:12.160000
 So that command would go here underneath
 your switched virtual interface.

0:10:12.160000 --> 0:10:27.780000
 And then the same logic would hold
 true as if it were a router.

0:10:27.780000 --> 0:10:33.220000
 Okay, lastly, what I want to talk about
 is configuring a router or a switch

0:10:33.220000 --> 0:10:36.240000
 as a DHCP server.

0:10:36.240000 --> 0:10:40.560000
 Like I mentioned earlier, this is perfectly
 fine for smaller environments.

0:10:40.560000 --> 0:10:42.440000
 There's a couple downsides though.


0:10:42.440000 --> 0:10:48.060000
 You see, the DHCP protocol has a ton
 of optional things you could do to

0:10:48.060000 --> 0:10:52.880000
 it. And not all of those options are
 available when you're configuring

0:10:52.880000 --> 0:10:55.520000
 a router or a switch
 as a DHCP server.

0:10:55.520000 --> 0:11:00.500000
 So if you want all the full set of features
 that DHCP can do, you probably

0:11:00.500000 --> 0:11:04.200000
 want to do that unlike a Linux server
 or a Windows server or something

0:11:04.200000 --> 0:11:09.240000
 like that. But just to get basic DHCP
 working, absolutely you could do

0:11:09.240000 --> 0:11:11.200000
 this on a router or a switch.

0:11:11.200000 --> 0:11:13.240000
 So here are the commands
 to do that.

0:11:13.240000 --> 0:11:17.780000
 Number one, at the global level,
 we'll want to type service DHCP.

0:11:17.780000 --> 0:11:20.340000
 I believe this is already
 on by default.

0:11:20.340000 --> 0:11:23.480000
 I've configured this in the lab environment
 before and I have purposely

0:11:23.480000 --> 0:11:28.820000
 not configured the command service
 DHCP and it still worked.

0:11:28.820000 --> 0:11:31.180000
 But just be aware that command
 works, that command.

0:11:31.180000 --> 0:11:35.780000
 If you go through your running config
 and you actually see no service

0:11:35.780000 --> 0:11:40.120000
 DHCP, well then there's no way this
 guy is going to be a DHCP server at

0:11:40.120000 --> 0:11:42.720000
 all. Now let's actually
 create the pool.

0:11:42.720000 --> 0:11:47.540000
 So at the global level, we say IP
 DHCP pool and they give it a name.

0:11:47.540000 --> 0:11:51.480000
 Whatever name you want, it's a descriptive
 name, it could be payroll.

0:11:51.480000 --> 0:11:57.200000
 Later on maybe we'll create a pool called
 HR, another pool called marketing.

0:11:57.200000 --> 0:12:02.060000
 And then within each pool, we're going
 to configure our basic bare bones

0:12:02.060000 --> 0:12:05.180000
 requirements of what
 our hosts need.

0:12:05.180000 --> 0:12:07.440000
 So we're going to
 configure network.

0:12:07.440000 --> 0:12:13.080000
 So for example, here,
 let's get rid of that.

0:12:13.080000 --> 0:12:16.280000
 Let's say we did this.

0:12:16.280000 --> 0:12:27.340000
 So I could configure my pool
 as IP DHCP pool INE.

0:12:27.340000 --> 0:12:29.920000
 So that pool is done.

0:12:29.920000 --> 0:12:44.380000
 Now within there, I could say
 network 20.110 slash IP.

0:12:44.380000 --> 0:12:48.600000
 Now notice this is one of those very
 few places where you can actually

0:12:48.600000 --> 0:12:53.900000
 type in a subnet mask as a slash
 notation, like slash 24 slash 29.

0:12:53.900000 --> 0:12:57.740000
 Now if you're used to typing it in
 as dotted decimal, you can do that

0:12:57.740000 --> 0:13:03.600000
 too. I could have typed network
 21.110 255.255.255.0.

0:13:03.600000 --> 0:13:08.180000
 So now it knows the range of possible
 addresses to pull from.

0:13:08.180000 --> 0:13:14.220000
 The very first address will clearly
 be 20.1.1.1 and the last usable host

0:13:14.220000 --> 0:13:19.860000
 address will be 20.1.2.54
 from that available pool.

0:13:19.860000 --> 0:13:21.520000
 And then we keep going.

0:13:21.520000 --> 0:13:25.180000
 These people are going to need a default
 gateway so we can configure that

0:13:25.180000 --> 0:13:33.340000
 as well. So like in my case, default
 dash router, maybe 20.1.1.

0:13:33.340000 --> 0:13:36.920000
 Maybe the very first router, the very
 first address on that subnet is

0:13:36.920000 --> 0:13:39.520000
 reserved for the router.

0:13:39.520000 --> 0:13:44.840000
 You can give a DNS server
 like, you know, 8.8.8.8.

0:13:44.840000 --> 0:13:47.580000
 That's a very popular
 one for Google.

0:13:47.580000 --> 0:13:50.960000
 And then you can give it a lease.

0:13:50.960000 --> 0:13:55.900000
 Now as far as the lease is concerned,
 if you leave the lease off, it does

0:13:55.900000 --> 0:13:57.340000
 have a default time.

0:13:57.340000 --> 0:14:02.560000
 So by default, the lease is going
 to be one day, 86,400 seconds.

0:14:02.560000 --> 0:14:05.560000
 But you can change that to anything
 you want by using this command right

0:14:05.560000 --> 0:14:08.320000
 here. And then that's it.

0:14:08.320000 --> 0:14:10.180000
 You've now created your pool.

0:14:10.180000 --> 0:14:13.920000
 Now one other thing.

0:14:13.920000 --> 0:14:19.520000
 There's a possibility that on this network,
 whatever this network is for

0:14:19.520000 --> 0:14:24.600000
 payroll, for example, there might already
 be some devices on that network

0:14:24.600000 --> 0:14:28.520000
 that have static pre-configured
 IP addresses.

0:14:28.520000 --> 0:14:32.260000
 For example, well, if we got a couple
 of servers on that exact same network

0:14:32.260000 --> 0:14:35.560000
 that payroll is going to use for, you know,
 that's where the payroll databases

0:14:35.560000 --> 0:14:37.140000
 are and things of that nature.

0:14:37.140000 --> 0:14:41.220000
 And those servers are going to be in the
 same range as your network command

0:14:41.220000 --> 0:14:46.940000
 right here. Well, we don't want DHCP
 to accidentally give a client the

0:14:46.940000 --> 0:14:51.300000
 exact same IP address that's currently
 configured on a static device.

0:14:51.300000 --> 0:14:53.180000
 We want to exclude that.

0:14:53.180000 --> 0:14:57.020000
 So that's what this is referring
 to right here.

0:14:57.020000 --> 0:15:00.820000
 Devices with static IP addresses, we
 need to exclude that from the pool

0:15:00.820000 --> 0:15:05.060000
 that's not done within
 the pool itself.

0:15:05.060000 --> 0:15:07.460000
 That's done at the global
 configuration level.

0:15:07.460000 --> 0:15:12.260000
 You can see that's the IP DTCP
 excluded dash address command.

0:15:12.260000 --> 0:15:15.620000
 If it's just one address, just
 type in that one address.

0:15:15.620000 --> 0:15:19.800000
 If you've got four or five devices in
 that network with static IPs, you

0:15:19.800000 --> 0:15:23.940000
 can do a range here with a starting
 and an ending IP address.

0:15:23.940000 --> 0:15:26.540000
 Now, how do we verify
 this actually worked?

0:15:26.540000 --> 0:15:30.620000
 Well, clearly, if it doesn't work, our
 clients are going to be very unhappy.

0:15:30.620000 --> 0:15:32.820000
 People are going to be calling us up saying,
 hey, I can't get on the network

0:15:32.820000 --> 0:15:33.700000
 what's going on.

0:15:33.700000 --> 0:15:37.440000
 And when you get on that client, you'll
 see, oh, it doesn't have an IP

0:15:37.440000 --> 0:15:40.920000
 address. So that would be your
 first clue that didn't work.

0:15:40.920000 --> 0:15:44.900000
 But on the actual router or switch
 itself, where you've configured all

0:15:44.900000 --> 0:15:48.580000
 these commands, there's two basic commands
 you can type in really one

0:15:48.580000 --> 0:15:54.660000
 command you can type in for verification,
 which is the show IP DTCP binding

0:15:54.660000 --> 0:15:59.240000
 command. So you can see right here,
 this clearly shows us that we've got

0:15:59.240000 --> 0:16:06.540000
 a couple of clients that have been allocated
 IP addresses, 10113 and 10111.

0:16:06.540000 --> 0:16:12.540000
 And we also see the client ID right
 there, which is unique per client.

0:16:12.540000 --> 0:16:18.900000
 And we can see when these addresses
 will expire, the least expiration.

0:16:18.900000 --> 0:16:23.600000
 So that concludes this video.

0:16:23.600000 --> 0:16:24.500000
 Thank you for watching.

