分类： 系统运维
2013-08-01 09:14:22
The ERX system supports s (IEEE 802.1q) on Fast Ernet and Gigabit Ernet interfaces. IEEE 802.1q describes two key Ethernet fields that have been added the Ethernet frame identify and prioritize traffic: ID (virtual LAN identifier) and user priority values (802.1p).
Ethernet frames using this standard can be tagged and viewed in the same way that ATM VCs (VPI/VCI) are represented on an ATM circuit. When an Ethernet frame is sent over a shared medium (for example, a GE line), it is logically defined as belonging to a specific virtual LAN and provisioned in the same way as a PVC on an ATM circuit. The Tag Protocol Identifier (TPID) in the Ethernet header indicates that this Ethernet frame contains 802.1q data. The VLAN ID indicates the virtual LAN (virtual circuit) to which this frame belongs. The three bits in the 802.1p field can be used to determine the priority of this frame.
On a given physical link, Ethernet frames may be tagged or untagged. The VLAN ID in the 802.1q tag explicitly identifies the VLAN for tagged frames; the ID may be 0. Untagged frames are implicitly in VLAN 0, which is considered to be the default VLAN. This implementation is different from ATM, where all frames are required to have an explicit VPI/VCI.
The VLAN ID field is 12 bits, allowing 4096 VLAN IDs. Tagged frames are identified by the value 0x8100 in the Ethernet protocol type field; the actual protocol type of the frame then follows the tag. Some attached devices may not accept 802.1q-tagged frames, and therefore can reside only in VLAN 0.
Conversely, some devices may accept only tagged frames, requiring that even frames in VLAN 0 be tagged. Ports that have been explicitly added to VLAN 0 will tag frames unless the untagged option is specified. Ports that do not have VLANs configured will not tag frames.
On ingress to a VLAN port, the Ethernet protocol type field of a frame is examined to determine if it is tagged. If so, the VLAN ID is extracted from the tag, and the actual protocol type field following the tag is used to identify the protocol. If not, the frame is in the default VLAN, and the type field is used to identify the protocol.
On egress, a frame destined for the default VLAN may be sent untagged or tagged (depending on whether the untagged option was specified). A frame destined for any other VLAN is sent tagged.
As specified in IEEE Standard 802.1q, the twelve-bit VLAN identifier's tagged frames enables the construction of a maximum of 4,096 distinct VLANs in an Ethernet B-RAS application environment; however, this VLAN limit is inadequate. A stacked VLAN (S-VLAN) provides a two-level VLAN tag structure, which extends the VLAN ID space to over 16 million VLANs.
Creating an S-VLAN requires the use of a second encapsulation tag. The ERX system performs decapsulation twice, once to get the S-VLAN tag and once to get the VLAN tag. Using this "double tagging" approach allows for over 16 million address possibilities, which more than satisfies the scaling requirement for Ethernet B-RAS applications.
VLAN and S-VLAN subinterfaces can coexist over the same VLAN major interface. S-VLANs and VLANs are configured similarly.
Ethernet interfaces are configured to operate in one of the following modes, depending on the combination of protocol and VLAN usage:
The Ethernet interface supports only one upper-layer protocol interface (without VLANs). Such support is backward compatible with the initial implementation of Ethernet interfaces.
The Ethernet interface supports a single upper-layer protocol on each of one or more VLANs.
The Ethernet interface supports multiple upper-layer protocols, one upper-layer interface per protocol, on each of one or more VLANs.
The first upper-layer interface configured on an Ethernet interface determines the mode. Switching to a different mode requires first removing the entire stack above the Ethernet interface. The supported upper-layer protocols are IP and PPPoE. shows the interface stacks for the different modes.
