Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

Guide That Contains This Content
[+] Expand All
[-] Collapse All

    Networking at Layer 2: Forwarding Dual-Tagged Frames

    The use of VLAN tagging to group (or bundle) sets of MAC addresses is a start toward a method of forwarding LAN traffic based on information found in the frame, not on IP address in the packet. However, there is a major limitation in trying to build forwarding tables based on VLAN tags. Simply put, there are not enough VLAN tags.

    Twelve bits only supply enough space for 4096 unique VLAN tags. This is hardly enough for all the LANs on a large corporate campus, let alone the whole world. A 12-bit tag might suffice for the local campus arena, but for the metropolitan area, comprising a whole city, more bits are needed.

    The number of bits in the VLAN tag, two bytes for the TPID and two bytes for the TCI field, are fixed and cannot be extended. However, another VLAN tag can be added to the frame, forming an inner and outer VLAN tag arrangement. This arrangement is defined in the IEEE 802.1ad specification and applies to devices that function on the provider bridge level. This means that Ethernet frames tagged at the local (or customer) VLAN level can receive another outer VLAN tag when they are sent to the provider's LAN switches. As a result, Ethernet frames can be switched across a metropolitan area, not just among the local organizations devices at the campus level.

    The outer tag defined in IEEE 802.1ad is often called the Virtual Metropolitan Area Network (VMAN) tag, a good way to recall the intended scope of the specification. The outer tag is placed after the MAC source address, moving the inner tag backwards in the frame. Both tags can be added at the same time by the same device (called a push/push operation), changed by a device (a swap operation), or removed by a device one at a time (pop) or together (pop/pop). Devices can perform elaborate variations on these operations (such as pop/swap/push) to accomplish the necessary networking tasks with the frames they process.

    The IEEE specification indicates that the outer tag of a doubly-tagged Ethernet frame should have a TPID value of 0x88a8. Any network device can easily tell if it has received a frame with one tag (0x8100) or two tags (0x88a8). However, because the value 0x8100 always means that a VLAN tag is present, most vendors and networks use the same TPID value (0x8100) for the inner and outer tags. As long as the configuration and processing are consistent, there is no confusion, and the TPID value can usually be changed if necessary.

    How do nested VLAN tags solve the VLAN numbering limitation? Taken together, the two VLAN tags can be thought of as providing 24 bits for tagging space: 12 bits at the outer level and 12 bits at the inner level. However, it is important to realize that the bits are not acted on as if they were all one tag. Even when the tags are nested, bridges on a provider backbone will normally only switch on the outer VLAN tag. All in all, the inner 12-bit tagging space is more than adequate for a Metro Ethernet network. Any limitations in the VLAN tag space can be addressed by adding more VLAN tags to the basic Ethernet frame.

    Modified: 2017-08-31