Overview

Configuring bandwidth limits for specific connected family devices on an Optical Network Terminal (ONT) is a critical requirement for B2B telecom engineers delivering managed residential and small-business services. Whether you are enforcing fair-use policies, prioritizing voice traffic, or capping guest Wi-Fi throughput, the ONT provides Layer 2/3 QoS mechanisms to shape per-device traffic. This FAQ addresses the most common pre-sales and post-sales technical questions about ONT bandwidth limiting, from optical module compatibility and maximum throughput to CLI configuration and redundant power failover.

Frequently Asked Questions

Q1: What is the maximum switching throughput and backplane capacity of a carrier-grade ONT for per-device bandwidth limiting?

The maximum switching capacity of a carrier-grade ONT is typically 10 Gbps to 40 Gbps on the backplane, with a packet forwarding rate of 14.88 Mpps for 10GbE line-rate traffic. This capacity is sufficient to enforce per-device bandwidth limits across 64 to 128 connected devices without introducing measurable latency. The backplane architecture uses a non-blocking fabric, ensuring that shaping one device does not degrade throughput for other subscribers on the same PON port.

Q2: Which optical module types are compatible with ONT models that support per-device bandwidth policing?

ONT models supporting per-device bandwidth policing are compatible with SFP/SFP+ optical modules for the uplink PON interface and 1000BASE-T or 2.5GBASE-T copper SFP modules for LAN-side expansion. Third-party modules must comply with MSA (Multi-Source Agreement) standards and the ONT's vendor-specific EEPROM whitelist. For XGS-PON deployments, use SFP+ OLT-compatible modules rated for 9.953 Gbps downstream and 9.953 Gbps upstream. Always verify the module's temperature range and digital diagnostic monitoring (DDM) support before deployment.

Q3: How do I configure per-device bandwidth limits on the ONT using the CLI?

To configure per-device bandwidth limits on the ONT via CLI, use the qos bandwidth-limit command within the interface or subscriber profile context. The basic sequence is: ont> enable, ont# configure terminal, ont(config)# subscriber-profile DEVICE_CAP, ont(config-profile)# bandwidth-limit downstream 100000 (in kbps), ont(config-profile)# bandwidth-limit upstream 50000, then apply the profile to the specific MAC address or DHCP lease. For per-device enforcement, bind the profile using ont(config)# mac-address-table static profile DEVICE_CAP. Verify with show qos bandwidth-limit interface.

Q4: What is the difference between per-device bandwidth limiting on current-generation ONTs versus previous generations?

Current-generation ONTs support hardware-accelerated per-device bandwidth limiting with 8 queues per subscriber and granular rates from 64 kbps to 10 Gbps, while previous generations relied on software-based rate limiting capped at 100 Mbps per device. The key difference is the use of a dedicated traffic manager ASIC that enforces shaping and policing at line rate without CPU intervention. Previous generations also lacked per-MAC scheduling and often applied limits globally per PON port rather than per connected device. Modern ONTs additionally support hierarchical QoS (HQoS) for parent-child shaping profiles.

Q5: Can I stack multiple ONTs to increase port density while maintaining per-device bandwidth limits?

Yes, multiple ONTs can be stacked using virtual chassis technology to increase port density while maintaining per-device bandwidth limits across the stack. Stacking interfaces typically use 10G SFP+ or 25G SFP28 ports in a ring or daisy-chain topology. The stacking backplane aggregates control plane and management traffic, but bandwidth limiting policies are enforced locally on each ONT's traffic manager. When stacking, ensure that the stacking links have sufficient headroom to carry inter-switch traffic without contending with subscriber-shaped flows. Use show stack status to verify ring integrity and show qos bandwidth-limit stack to confirm policy synchronization.

Q6: How does redundant power failover affect per-device bandwidth limiting on the ONT?

Redundant power failover on the ONT does not affect per-device bandwidth limiting because QoS policies are stored in non-volatile memory and re-applied automatically upon power restoration. The ONT supports dual AC or DC power supplies with automatic failover in under 10 ms, ensuring that traffic shaping continues uninterrupted during a single power supply failure. If both supplies fail, the ONT reboots and reloads the saved configuration, including all bandwidth profiles, from flash. For carrier-grade deployments, use DC power with a central office battery plant to achieve 99.999% availability. Troubleshoot failover issues with show power-supply status and show qos bandwidth-limit active.

Q7: What QoS mechanisms does the ONT use to prioritize time-sensitive traffic while enforcing bandwidth caps?

The ONT uses 802.1p priority queuing, DSCP marking, and weighted fair queuing (WFQ) to prioritize time-sensitive traffic while enforcing bandwidth caps. Each subscriber profile can map voice (EF), video (AF41), and best-effort (BE) traffic to separate queues with guaranteed minimum rates and shaped maximum rates. This ensures that a bandwidth-limited device cannot starve VoIP or IPTV streams. The ONT also supports strict priority queuing for control-plane protocols such as OSPF and VRRP. Use show qos queue-statistics to monitor queue depth and drop counters.

Q8: How do I troubleshoot per-device bandwidth limits that are not being enforced correctly?

To troubleshoot per-device bandwidth limits that are not enforced correctly, first verify that the subscriber profile is bound to the correct MAC address or DHCP lease using show subscriber-profile binding. Common causes include incorrect profile assignment, MAC address randomization on client devices, or the device being behind a NAT router that hides its MAC. Second, check that the traffic manager ASIC is not in bypass mode with show qos status. Third, confirm that the bandwidth-limit values are in the correct units (kbps vs. Mbps). Finally, capture traffic with monitor capture interface to verify actual throughput against the configured cap. If the limit is applied but not enforced, a firmware upgrade may be required to address known QoS bugs.