Hello, my name is Tyler McMinn with Aruba Networks. This is part 3 of our networking essentials series of videos. In the last video, we had rebuilt our entire part 2 scenario with everything working until we tried to route between two different subnets. In this video, we're going to address the issue of routing. Sit back and let's get started. Routing and gateways, aka the great escape. What routing is, is the ability to jump between one interface on a layer 3 device, one subnet, or street, if you will, of devices that can all talk to each other and another subnet or another interface on a router where all of these devices can all reach and talk to each other. In the example they're showing VLAN20 on one side, VLAN10 on the other. The VLANs don't really matter at this point, what's imperative here is that you have a layer 3 function in the middle. I'll draw that in green here. That layer 3 router, every interface must be in a different subnet. It won't let you apply the same subnet on two different interfaces on this router. In the design of the network, we've got host A wanting to speak with host B and because they're on different interfaces of this router, they are necessarily in different subnets. Therefore, they need to get to a gateway. They need to get to their assigned gateway for their particular subnet in order to have reachability out of their network. This is going to be their default gateway address for router A or for PC A here, it's going to be this interface on the router that'll be its default gateway. What the router will do, is it will build a forwarding table. Much like a switch does forwarding between a layer 2 MAC address and a layer 1 interface, a router is going to carry a routing table or maintain a routing table with paths to layer 3 destinations and what layer 2 or layer 1 interface it's going to use to get out or in some cases we would call this the next hop address of where it's going to send it. Effectively it just needs to know where to forward the frames or to forward the packets on your behalf. That is the design or the goal of routing is to essentially route your traffic. Routing sits at layer 3 of our OSI model here. You probably remember that from our part one discussion where we went through these seven layers of the OSI model. Seventh layer just being the application layer, which is oddly absent here. But at layer 3, we have a source IP address and a destination IP address. These are IPv version four addresses that are represented in dotted decimal to show 32-bits of information. IPv6 would be in hexadecimal and would show 128 bits of information, so much larger addresses. For this discussion, we're just going to stick with IPv4 as it's still by far the more popular protocol used out there. To route these layer 3 addresses to a particular destination and that's primarily what we use is we use the destination IP address. We need some device that is smart enough or technical enough that it can look at the layer 3 header and make a forwarding decision using and maintaining a routing table. Now routers are traditional devices that that's what they do, that's their dedicated function. But more and more the use of switches with a lot more ports, a lot more flexibility and usually less expensive, would be able to run initially some basic routing functions and more advanced routing functions like here 8,400 is arguably a CX multilayer switch, but can support over a million BGP routes. It can route the entire world's Internet forwarding table or forwarding routing table. Not lacking in the feature department at all. I gave a quick little roundup on the discussion between IPv4 and IPv6. Don't let this throw you off with IPv6. It's actually a lot easier than it looks with these crazy along addresses that they're showing you here. But even with IPv4, it was developed by the ARPANET project in the late 60s and back then, the idea of having over 4 billion public IP addresses being used was unheard of. It was thought that there would be no more than maybe a few thousand mainframe computers in the world ever, that there would never be more than a need for that. The idea of a 32-bit address was something that they would never ever need to change. But we did discover through the 80s and really in the 90s that we were running a bit short of addresses, and they started development on a replacement protocol for that. Our computers nowadays support both. You'll find that you're running both IPv4 and IPv6. Your router support both, all of our CX gear, all of our Roombas switches supported IPv6 and IPv6 protocols for not just public or static routing with IPv6, but also OSPF version 3 if you wanted to support that as well and multiprotocol BGP. They were completely prepared for a full transition. We support customers that use IPv6. But we generally don't teach it very much simply because of a lack of demand for IPv6 as of yet. The IPv4 address is 32-bits. Breaking these ones and zeros down is great although your switches and your computers and your network interface cards, they understand the binary very well. Us as humans, we need to broken up into these octets here of eight bits each in order to make it a little more bite size and palatable for us to understand. Each eight bits is called an octet, and then that can be shorthanded in decimal format. We don't need to get into the binary piece of it, but that's always a good skill to have. We dive into that in the five-day fundamentals class. Ultimately, these octet representations constitute the four octets of any IPv4 address. Public, private doesn't matter, all addresses are 32-bits, they're just represented uniquely using an IP address. The mask applies to that address to break it up into two parts. The first part on the left is the network identifier. It's the street or the subnet that I've been referring to, the network that you belong to. The right-hand side or the bits that are masked by zeros is known as the host portion. Where you draw that line between the host identifier and the bits that represent the network identifier, that's basically the job of a subnet mask is to draw that line where the bits to the left are the bits that represent the network you belong to, and everybody in your subnet, everybody in your little local subnet belong to, and your bits to the right represent their individual host addresses, so there's no duplicates within that subnet. When you leave your subnet that's where routing takes place, you require a router in order to get to that other destination. A quick pop quiz, we're not going to do a separate video I'll just leave it in here. Given an IP address of 172.20.3.54 and a mask of 255.255.255.0, what can be accurately stated about this addressing? There are three answers here. Now, this is a little bit more difficult if you haven't worked with IP addresses before, but I think I've answered all the questions, so take a moment, read through the options, pause the video. All right, are you back? Do you think you know the three answers? Let's take a look at what the answers are. B, the network portion of the address is 172.20.3. They're saying that this is the network portion and that is correct because it's underscored by the three octets of 255s. If it had a zero, that would be the host portion, which is to say the host 172.20.3.89, would be on the same network as this guy who has a host address of dot 54, this would be dot 89. It's a unique address, but that doesn't make it on the same network. What makes it on the same network or subnet would be that the first three octets, or the first 24-bits slash 24 or these octets of eight bits plus eight bits plus eight bits plus zero for the host side, these 24-bits are telling us what network that it belongs to. Yeah, that is correct. Any host range from dot one through dot 254 would be a valid host within that range. Two hundred fifty-five is a reserved address and dot zero it's used for the identifier portion of it. You can only go as high as 255, that's with all the bits flagged, so there is no 256. The final one, I highlighted that as well, you could also indicate the mask as slash 24, because each octet that's flagged with all of its bits as ones adds up to a value of 255, not allowing anything higher than that, so no 256. Those are all correct. What is wrong? The host portion of the address is three and 54. No, the line is drawn here where the 255s end and the zero begins, so this is the network portion, to the right is the host portion. No, this is incorrect. Then D, a switch might use this address to forward the packet. I think what they were getting at here, given that this is supposed to be wrong. If you had to assign this as a switch management address, sure, you could use this as a management, that's fine, it's a valid address, but would you use this address to forward the packet? Would a router, I think that's what they meant to say. If you got this one wrong, I apologize. If you wrote it as a router forwarding to forward the packet or a multi-layer switch would make a little more sense here. Multi-layer switch, yes. You would actually not use this to forward the packet, it doesn't represent the network identifier. The network identifier would be the very first address and the entire scope, 172.20.3.0 is the lowest number in that range, and the last number would be dot 255, both are reserved. This is the network identifier, this is what we call the local broadcast or directed broadcast. Both of those you wouldn't use to assign to hosts themselves, like the switch itself as an address or a PC, instead you would use dot one through dot 254. Yeah, it's a lot of math when you get into the subnetting and that portion of it, so we just wanted to touch on that to get a better understanding of addressing. Let's go ahead and stop the video here. In the next video, what we're going to do is, we're going to cover inter-VLAN routing with multi-layers switches, describe what's going on there. We've already actually talked about that a little bit and then that'll bring us to our first lab where we'll start building this out. Again, thank you for your time on this, I'll see you guys in the next video.