Subtitles of the Movie

This module takes a look at DHCP in a routed environment. If you recall from our previous discussion of the DHCP lease process those 4 packets are broadcast based. Now what do you do if you are in a routed environment and the most networks are, even a office building between rooms typically now a days has a router in between. What do you do to implement DHCP when our routers won't pass these broadcast packets? Well maybe they can. One thing you can do, one of your options and you have basically 3 is that you can install a 1542 compliant router. So if here was our simple routed network. We could, if this was 1542 compliant, then this router listens for BOOTP traffic, which the discover, the offer, the request, the acknowledge all are, and then it forwards these to other subnets. Would these guys have a problem getting an address from DHCP server? No. They don't have to go through the router. But these guys do. So to go through the router this router, this 1542 compliant router will forward BOOTP protocol traffic. So that's one thing you can do and most modern routers do support this. The other thing you can do is you can configure a DHCP relay agent and when you do that you will configure, and we'll look at this actually in next module, you'll designate one of the computers as a relay agent and this in turn, and this won't exist because you wouldn't implement these both at same time, but this would listen for those discover packets and this would be configured with the IP address of the DHCP server. So it can take discover messages and route them directly or send directed IP traffic through the router to the DHCP server. So that's another thing you can do, the relay agent access a proxy on behalf of your DHCP clients. Your third option is that you configure a DHCP server on each subnet. So now we don't use this, we're not using a relay agent, we just make this server a DHCP server as well. Now we have a router that does not pass BOOTP traffic and these clients get their IP addresses from this server, these clients get their IP addresses from this server. So those are the 3 options that you have. Now if you're using, if 1542 compliant router or DHCP relay agent one of these solutions especially 1542 compliant router, you can still implement two DHCP servers for fault tolerance and backup. Let's pretend that this is a 1542 router so it passes BOOTP traffic but we still have two DHCP servers in the network, and this becomes something that is ripe for test questions by the way, and the reason you have these two DHCP servers is that you want one to be the backup for the other in case there is a failure in the networks somewhere. What you want to configure, what you want to use is the 80 20 rule and what the 80 20 rule specifies is that if this is subnet1 and this is subnet2, then when you create your scopes you create out of the pool of subnet1's addresses, you have 80% from subnet1's in that pool here, and then over here you have 20% of the available subnet1's IP addresses here. And then for subnet2 20% of subnet2's available addresses and 80% of subnet2's. These clients you want them to get their IP addresses from this pool right here. These clients you want them to get addresses from this pool here. But in case this DHCP server goes down, then these clients have to go through the router and pull IP addresses from here. When they do you don't want them to get IP addresses from this range here because this subnet will have a different default gateway as this subnet, and so these computers wont be able to communicate. So you want these computers to grab an IP address from this pool here and that way you can implement backup and fault tolerance and still have a working network. Now the question often becomes how does it know? How does this DHCP server know to offer an IP address from this range rather than this range? Well when the discover packet is passed through the router, the router then appends or changes the source IP address with the network interface card, the IP address of the router interface here. So once that packet hits the DHCP server, the DHCP server looks at the contents of the packet, notices that it's coming from this interface card, and then says okay, well that interface card is on subnet1 so let's offer an IP address from subnet1 and so that's how it all works and that's how it would work with a relay agent as well. So again we have 3 choices for implementing DHCP in a routed environments or we implement maybe just one DHCP server. And typically that's all you need. Install a 1542 compliant router, configure DHCP relay agent or however many subnets you have, configure 1 DHCP server for every subnet. If you use a fault tolerant implementation remember use the 80 20 rule to break up your scopes.