Overview

QinQ (802.1ad), also known as Stacked VLANs or VLAN double-tagging, is a critical technology for service providers and large enterprise campuses. It allows operators to encapsulate a customer's inner 802.1Q VLAN tag within an outer service-provider tag, preserving customer VLANs across a metro Ethernet core. This FAQ addresses the most common pre-sales and post-sales technical support questions, from optical module compatibility and maximum throughput to CLI configuration and redundancy troubleshooting.

Frequently Asked Questions

Q1: What is the maximum switching throughput and backplane capacity of a typical QinQ-enabled aggregation switch?

The maximum switching capacity for a QinQ-enabled carrier Ethernet switch ranges from 176 Gbps (for 24xGE + 4x10G models) to 2.56 Tbps (for modular chassis with 10GE/40GE uplinks). Forwarding performance typically scales to 132 Mpps for full line-rate forwarding on all ports. Because QinQ adds a 4-byte outer tag, the switch must support jumbo frames (at least 1522 bytes, ideally 9216 bytes) to avoid packet drops. The backplane should be non-blocking; always verify that the vendor specifies 100% throughput with double-tagged frames.

Q2: Is QinQ backwards compatible with standard 802.1Q trunks, and what is the main difference from VLAN mapping?

Yes, QinQ is fully backwards compatible with standard 802.1Q trunks. The main difference is that QinQ adds an outer tag (S-TAG) while preserving the inner customer tag (C-TAG). VLAN mapping, by contrast, replaces or translates the customer VLAN ID entirely rather than encapsulating it. QinQ preserves end-to-end customer VLAN transparency, making it superior for multi-tenant environments. However, legacy switches that do not support jumbo frames may drop QinQ frames (1522 bytes vs. standard 1518 bytes).

Q3: What is the basic CLI configuration to enable QinQ on a provider-edge port?

The standard approach is: 1) Set the port to trunk mode: 'switchport mode trunk'. 2) Define the outer S-VLAN: 'switchport trunk native vlan tag' (on some platforms) or 'vlan dot1q tag native'. 3) Enable QinQ encapsulation globally: 'vlan dot1q tunnel' or 'switchport mode dot1q-tunnel' on the customer-facing port. 4) Map the customer edge VLAN to the service VLAN using 'switchport access vlan ' combined with 'switchport mode dot1q-tunnel'. Note that TPID must often be set to 0x88a8 (standard 802.1ad) vs. the default 0x8100 (802.1Q).

Q4: What optical module SFP/SFP+ compatibility constraints exist for QinQ uplinks?

QinQ itself does not impose optical constraints, but the uplink optics must support the maximum frame size (jumbo frames) and the data rate of the outer S-VLAN. For 1GE QinQ trunks, standard SFP 1000BASE-T or SFP Gbic (1310nm SM) work. For 10GE aggregation, SFP+ SR (300m OM3) or LR (10km SM) modules are required; ensure the PHY chipset supports 9728-byte MTU. Third-party optics are compatible if the switch does not enforce coded EEPROM vendor locking. Critical: For 40GE/100GE QinQ core uplinks, use QSFP+ LR4 or QSFP28 modules; the outer tag processing on high-speed interfaces must be hardware-accelerated, not CPU-processed.

Q5: How do I troubleshoot QinQ packet loss or MTU-related drops across a stacked virtual chassis?

First, verify MTU consistency across all stack members using 'show interfaces mtu'. QinQ requires end-to-end MTU ≥ 1522 bytes; enterprise standard is 9216 bytes. Second, check for TPID mismatches using 'show vlan dot1q-tunnel'. Third, test with ICMP jumbo frames: 'ping size 1524 df-bit'. For stacking errors, ensure that the stack ring supports jumbo frame forwarding; some stack interconnects default to 1500-byte MTU. Finally, use embedded packet capture or SPAN to verify double-tagged frames exit the egress port correctly. Common post-sales fix: 'system mtu 9216' reload required on many platforms.

Q6: Does QinQ support L3 protocols like OSPF or BGP on the outer S-VLAN interface?

Yes. On QinQ termination nodes (PE routers), you can configure an L3 VLAN interface (SVI) on the outer S-VLAN ID. For example, 'interface vlan 1000' (where 1000 is the S-VLAN) can run OSPF or BGP with neighbor service providers. However, QinQ inner tags remain transparent to L3. Warning: Control plane protocols should run on untagged or single-tagged interfaces; double-tagged L3 termination requires specialized hardware (ASICs that pop both tags before L3 lookup). Most carrier switches support S-VLAN L3 termination, but performance degrades if C-VLANs are also used for L3 routing.

Q7: How does redundant power supply failover work for QinQ aggregation switches in a live metro network?

In a redundant AC/DC configuration, QinQ switching continues uninterrupted during PSU failover if the switch uses hitless power redundancy (1+1 mode). The switching ASIC and PHYs remain powered by the secondary PSU within 10-50ms. Troubleshooting steps: 1) Monitor 'show power' to confirm both PSUs are 'active/standby'. 2) Simulate failover with 'redundancy force-switchover' during off-hours. 3) Check syslog for 'PSU failure' traps. Critical post-sales note: Some low-cost switches reboot all QinQ forwarding tables upon PSU reinsertion, causing traffic loss. Verify that your vendor supports hot-swappable PSUs with stateful failover (sub-100ms recovery).

Q8: What is the maximum number of QinQ VLAN translations supported per port or per system?

Typically, enterprise QinQ switches support up to 4094 outer S-VLANs and up to 4094 inner C-VLANs per S-VLAN (total 16 million combinations). However, hardware TCAM limits practical scale. A typical Broadcom Trident2-based switch supports 32,000 QinQ translation entries. For high-density deployments, verify the exact VLAN translation table size using 'show vlan translation table' or 'show hardware capacity'. Exceeding this limit forces packets to be software-processed, dropping throughput below 1 Gbps. Always calculate required QinQ scale pre-sales: (Number of S-VLANs) × (Average C-VLANs per customer).