Tech Brief Video Series – VXLAN

Hello, I have created a “VXLAN Tech Brief” Series. Kindly click on the links to view the videos. I’m trying out a few videos styles.

VXLAN A – VXLAN – Virtualization & VM Mobility – >

“The traditional physical world consists of single server application deployments where one application is installed on a single physical server. This era proved an inefficient use of server hardware resource as single app per server deployments never fully utilized hardware resources such as RAM and CPU. Then along came server virtualization and the hypervisor. Server virtualization provides the ability to carve a physical server into multiple hosts, known as Virtual Machines (VM). Each VM operates with its independent Operating System/Application utilizing an abstraction layer of the physical host’s hardware – CPU, RAM, and NIC. The server’s hardware is shared among multiple VMs.”

VXLAN B – VXLAN – Dynamic MAC Learning – >

“Initially, Ethernet started with a Thick Coax Cable – a single cable was used to connect all workstations together. It was later replaced by the twisted pair cables – unshielded twisted pair UTP and shielded twisted pair STP in the late 1990 – 2000’s. On Ethernet networks, each host has a unique MAC address for identification. Devices with multiple NICs require multiple MAC addresses, one MAC address per NIC. So how do you talk to a group of hosts and ensure that all other hosts don’t receive the traffic?

Broadcast and Multicast communication. Multicast involves sending to a group of receivers in a single stream, whereas broadcast involves sending to all receivers. Multicast is similar to a broadcast in the sense that it sends to a group of machines. However, multicast sends to SOME MAC addresses while broadcasts send to ALL MAC addresses. All hosts receive broadcast traffic whether they like it or not. When a host sends a packet to the broadcast MAC address (FF:FF:FF:FF:FF:FF), it is delivered to all stations on the wire. A Broadcast received by a host requires processing so it’s a good idea to keep broadcasts to a minimum on your network.”

VXLAN C – VXLAN – Introducing Overlay Networking ->

“Essentially, an overlay is either placing Layer 2 or Layer 3 over a Layer 3 Core. The Layer 3 Core is known as the underlay. This removes a lot of the drawbacks and scaling issues which we had with traditional Layer 2 connectivity which uses VLANs. The multi-tenant nature of overlays is designed to get away from these L2 challenges allowing you to build networks at a much larger scale. Layer 2 and layer 3 overlays. We have both Layer 2 and Layer 3 overlays. Layer 2 overlays emulate a Layer 2 network and map Layer 2 frames into an IP underlay. If you are emulating a Layer 2 network, you have to somehow emulate Layer 2 flooding behavior. This is the bread and butter how Layer 2 networks work, and that doesn’t change just because you decide to create a Layer 2 overlay.”

VXLAN D – VXLAN – Introducing VXLAN – >

“The VLAN tag field defined in 1. IEEE 802.1Q has 12 bits for host identification, supporting a maximum of only 4094 VLANs. Its common these days to have a multi-tiered application deployment where every tier requires its own segment and with literally thousands of multi-tier application segments this will for sure run out. Then came along Virtual extensible local area network (VXLAN). VXLAN uses a 24-bit network segment ID, called a VXLAN network identifier (VNI), used for identification. This is much larger than the 12 bits used for traditional VLAN identification.”

VXLAN E – VXLAN – VXLAN Operations – >

“Finding out the destination VTEP. The big decision is how you discover the destination VTEP IP address. The destination VTEP IP address needs to be mapped to the end host destination MAC address. The mechanism used to do this affects the scalability & VXLAN domain functionality. We need some kind of control plane element. The control plane element of VXLAN can be deployed as a flood and learn mechanism, which is not a real control plane or you can have an actual control plane (that does not flood and learn) or even use an orchestration tool for VTEP to IP mapping. Many vendors implement this differently.”

VXLAN F – VXLAN – VXLAN Phases – >

“VXLAN went through a number of phases for ways to get the remote VTEP IP information. Initially, it started with a flood and learn the process and finally moved to a use a proper control plane – EVPN. EVPN is a pretty good control plan as previous methods rely on data plane Flood and Learn behavior hindering the scalability of VXLAN domains.”

More videos to come!

Additional VXLAN information can be found at the following:

About Matt Conran

Matt Conran has created 184 entries.

Leave a Reply