How VXLAN Works Example

So what is VX Land? VX Lands technology that allows you to extend the layer to overlay a free segment typically used in a data center. It uses all the bandwidth of the uplinks because the uplinks have no layers, so they will not be blocked by spanning tree. gives you the flexibility to extend layer two, not in the typical trunking way over multiple bits of hardware, it relies on two tunnel endpoints, so your switches at the endpoints need to know about VX LAN , while all these things in between don't need a Don't worry, in this video I will give you a high level

example

of how a host contacts another host using VX LAN.

If you're seeing more videos like this, don't forget to subscribe to put this in context. I'm going to start with a VLAN

example

. I have two hosts, two switches, one connects to the switch and those two can switch and there is a trunk between both switches. Everything is on the same VLAN. Both PCs are turned on. Both switches do not. I don't know about any of the PCs. Host one tries to communicate with host two and sends a frame to change a switch. One receives the plot. Put it in the source MAC address of host one in your MAC address table and don't see it. anything for a host two, so it sends a broadcast that floods across all ports that will fall into the same VLAN of host one.

Switch two receives the transmission from switch one. It inspects the MAC address and places it in its MAC address table which is the host. Switch two then makes a broadcast from all of its ports that belong to the same VLAN and host two responds by sending the frame back to switch two, which then places the source MAC address of host two in its MAC address table and forwards it. . To change a switch, one receives the response and places the source MAC address of host two into its MAC address table. You can now forward the frame that host one initially sent, so that's an example of how two hosts communicate with each other on the same VLAN.

For example, because the To say that that allows you to have more functions, which is the former ground, we are going to remove the trunk between those switches and now we are going to connect them to a spinal switch that is essentially a free device that is capable of meeting our needs, the staff from VX Land, everything happens on the leaf switches, the backbone actually doesn't need to know anything about VX Land, putting a disclaimer here, that is, if you don't use any VX Land on the backbone, we need to put some configuration initial before we can use VX land, so to start we need to assign IP addresses to our uplinks, we need to push some loopback addresses into the devices, we need to make sure there is routing on all those IP addresses so we can use OSPF. or even static routes that will create our underlying network, our simple underlying network is almost configured.

The next thing we need to think about is how the transmission overcomes these layer-free links. The way VX Land does this is with two methods, one is to use multicast. the other is to use Ingress for replication, it's one or the other and in this example we will use multicast with Ingress replication, what it does is it uses the broadcast framework, it replicates that into unicast and sends it to all the other VX grounds. tunnel endpoints, so it creates more bandwidth on the links, which is not the best thing to do, but sometimes people can't introduce multicast over their network, the X LAN will use multicast To get around the restriction that Layer 2 gives us with surge traffic, what is transmitted unknown unicast and multicast traffic in our previous example, where switch one did not know the destination MAC address for host 2, created a broadcast .

This is exactly the same with X land, however, that switch instead of flooding the trunk port, it floods it. its V point, which is the endpoint of the VX ground tunnel, that V point has the ability to take that transmission and propagate the information using multicast, so we need to configure multicast through our column and leaf switches, being the column the underlying RP. The net

works

are now configured so we can introduce the VX LAN. First, it assigns a VLAN to a V ni v and is essentially the excellent number, so we can assign VLAN 10 to vni ten thousand ten.

That ten thousand ten number is globally significant, so all your switches will have to share that if they are going to be on the same VX LAN, the local significant number of VLAN 10 can be different on any of the switches, it is the VX number LAN v nor you need to worry about. What I mean by that is that VLAN 10 could be assigned to VNA ten thousand ten on switch one on switch two, it could be VLAN 20 assigned to VNA ten thousand ten because they both belong to V and I ten thousand ten, they both can communicate. each other, it doesn't matter that there are different VLANs, it shows you that it can be done, but it is probably not best to mix and match.

If we can help you with VLAN to full vni mapping, now we can configure the VX ground tunnel endpoint. Known as type V, type V is the exit point for the local switch of a VX LAN to reach the same LAN VX that is on a remote switch, so we essentially have to tunnel between the endpoints and a tunnel that connects those endpoints to each other, what is known as the Overlay Network is the same concept as an IPSec tunnel, for example, if you have two firewalls with an IPSec tunnel through the Internet, the Internet is the young delay, the IPSec tunnel is the default overlay, VX LAN tunnels are not encrypted, however you can choose to encrypt them, each of the VN eyes must become a member of the V tab, once you remember it, they must join a group multicast on each of the V and ice, if v ni 10010 joins the multicast group at 39.1 point 1.1, then that must be in place on each leaf switch that has v ni 10010 in this is so that transmissions can be propagated to each of them via nice, each of the different VN eyes can join the same multicast group, this is because when the V point receives a multicast packet, it inspects the vni of the multicast packet and decides whether to propagate it if it matches all or discard it with this instead.

Now we can give an example of host one indicating host over this network. Host one sends a frame to change a switch, host one adds the source MAC address of host one to its MAC address table. The switch does not know the MAC address of host two, so it creates a broadcast broadcast that is forwarded from all ports that belonged to the VLAN of that host once it is in that veena and is assigned to VN I 10010 BNI 10010 es member of the V tip, so the transmission is forwarded from that V tip, the V tip encapsulates that frame with a VX LAN header UDP header and the IP header also an Altima.

Killer, this is then forwarded for a multicast to the RP, the RP sends a copy to each of the V taps that have joined the multicast group. In our case, change two V tips. There is additional payload to consider here, so increasing the MTU is necessary. o decreased on the hosts switch to receive the multicast packet on its tab V inspects V and I and has the member so D caps the packet now has the initial frame that was sent from the switch one inspects the frame for the MAC source address and assigns it to its V tab whose - then responds to this broadcast when it switches to sees the response, adds the MAC address to its MAC address table for that interface and forwards the V tab of the responding virus before it is forwarded, it is encapsulated again with the VX LAN header, the UDP header and an IP header, the destination IP address will be the V tip of switch one and the source IP address will be the V tip of switch two because we configured the routing correctly in the underlying network, the underlying is able to allow this traffic change, one then receives the packet, the D capsule reviews it and inspects the response and can add the MAC address of the two to its MAC address table, assigning it to the V tab, the communication from host 1 and host 2 can then be done for a unicast, when the unicast between host 1 and host two is sent, it goes through VX LAN encapsulation and decapsulation, so we still have We have to use flooding and learn how to propagate our mac address tables, which is less acceptable in a small environment like this example.

When we start adding many devices, flooding and learning can become more and more frequent to reduce flooding. We know we can use bgp evpn. Bgp evpn allows switches to share the information they learn from the flooding and learning process. This allows this solution to become much more scalable and much more effective by allowing each switch to propagate information that they know in their MAC address table, which in turn makes this solution much more scalable. They have been a little too detailed on some points of this but I think this is As high a level as I can, I hope the videos have helped you and if you find them useful please give me a thumbs up if you want to see more videos like this please subscribe to the channel if you want to contact me, leave a comment in the section below or contact me at Aspire Bank on social media, Lewis Burbank of our website.

Don't worry, check it out and thanks for watching.