Virtual Extensible LAN
Network tunneling protocol
From Wikipedia, the free encyclopedia
Virtual eXtensible LAN (VXLAN) is a network virtualization technology that uses a VLAN-like encapsulation technique to encapsulate OSI layer 2 Ethernet frames within layer 4 UDP datagrams, using 4789 as the default IANA-assigned destination UDP port number,[1] although many implementations that predate the IANA assignment use port 8472. VXLAN attempts to address the scalability problems associated with large cloud computing deployments.[2] VXLAN endpoints, which terminate VXLAN tunnels and may be either virtual or physical switch ports, are known as VXLAN tunnel endpoints (VTEPs).[3][4]
History
VXLAN addresses the limitations of the VLAN standard for segmenting a physical network into multiple virtual networks. The 12-bit VLAN identifier limits the number of virtual networks on a physical network to 4096, while in some applications, IEEE 802.1ad can be used to extend this limit, this is generally insufficient for large-scale cloud computing infrastructures with many tenants. Also, the VXLAN standard provides for extending a virtual network across multiple geographically distributed physical networks by encapsulating layer 2 frames in UDP packets. This is useful for large distributed data centers.[5]
VXLAN is an evolution of efforts to standardize on an overlay encapsulation protocol that addresses these issues. Compared to single-tagged IEEE 802.1Q VLANs, which provide a limited number of layer-2 VLANs (4094, using a 12-bit VLAN ID), VXLAN increases scalability up to about 16 million logical networks (using a 24-bit VNID) and allows for layer-2 adjacency across IP networks. Multicast or unicast with head-end replication (HER) is used to flood Broadcast, unknown-unicast and multicast traffic.[6]
The VXLAN specification was originally created by VMware, Arista Networks and Cisco.[7][8]
Implementations
VXLAN is widely, but not universally, implemented in commercial networking equipment. Several open-source implementations of VXLAN also exist,[9] including built-in capabilities in the Linux kernel.[10]
Commercial
Arista, Cisco, and VMware were the originators of VXLAN and support it in various products.
Other backers of the VXLAN technology include Adtran,[11] Arrcus, Big Switch Networks, Broadcom, Citrix, Cumulus Networks, Dell EMC, Ericsson, Huawei,[12] Joyent, Juniper Networks, Mellanox,[13] MikroTik,[14] Netgate, Netgear, [15] Pica8, and Red Hat.[16]
Open source
- FreeBSD,[17]
- OpenBSD,[18]
- Linux,[19]
- Open vSwitch is an example of a software-based virtual network switch that supports VXLAN overlay networks.
Standards specifications
VXLAN is officially documented by the IETF in RFC 7348.[16] VXLAN encapsulates a MAC frame in a UDP datagram for transport across an IP network,[20] creating an overlay network or tunnel.
Alternative technologies
Alternative technologies addressing the same or similar operational concerns include:
- IEEE 802.1ad ("Q-in-Q"), which greatly increases the number of VLANs supported by standard IEEE 802 Ethernet beyond 4K.
- IEEE 802.1ah ("MAC-in-MAC"), which supports tunneling Ethernet in a way that greatly increases the number of VLANs supported while avoiding a large increase in the size of the MAC Address table in a Carrier Ethernet deployment.
- Network Virtualization using Generic Route Encapsulation (NVGRE), which uses different framing but has similar goals to VxLAN.
See also
- Distributed Overlay Virtual Ethernet (DOVE)
- Ethernet VPN (EVPN)
- GENEVE, an industry effort to unify both VXLAN and NVGRE technologies
- Generic routing encapsulation (GRE)
- IEEE 802.1ad, an Ethernet networking standard, also known as provider bridging, Stacked VLANs, or simply Q-in-Q.
- IEEE 802.1ah, an IEEE Ethernet networking standard, also known as Provider Backbone Bridging (PBB) or MAC-in-MAC.
- NVGRE, Network Virtualization using GRE, which is a similar competing specification to VxLAN.
- Overlay Transport Virtualization (OTV)
- Virtual LAN (VLAN)
- Layer 2 Tunneling Protocol (L2TP)