Unlock full network capacity with intelligent shaping.

The answer depends on how the access point performs during peak hours, not on its nominal capacity, so the assessment should start with measured data. First, look at peak-hour utilization: the combined throughput of the subscribers served by the access point compared with the capacity it actually delivers in its radio conditions. Second, look at subscriber experience during those same hours: latency, packet losses, retransmissions and congestion. An access point with high utilization but low latency and losses still has headroom, one where latency and losses rise sharply at peak is already at its limit.

Bequant provides this data per subscriber and per subscriber group, so an ISP can group subscribers by access point and compare peak-hour throughput with QoE metrics. It provides an automatic list of the worst percentile of congestion-suffering subscribers, showing which access points are under pressure.

Additional load should then be validated, not assumed: add subscribers gradually and confirm that peak-hour latency, losses and congestion remain within target. Bequant congestion mitigation and AQM can increase the headroom of an existing access point by keeping latency and losses low as it approaches saturation. The reported gain is, for example, around 10% more users per access point, but it depends on each network, its radio conditions and its traffic mix. A Bequant free trial is the best way to measure it on your own network.

Application-based shaping uses deep packet inspection (DPI) to identify the applications generating the traffic and to apply policies to specific ones, for example limiting video streaming or large downloads. It is the right approach when a few bandwidth-hungry applications drive most of the peak demand and moderating them frees capacity for everything else.

Topology-based shaping applies a rate limit to the combined traffic of all subscribers served by a shared element, such as an access point, a sector, an OLT port or a backhaul link, matching the capacity of that element and sharing it fairly among its subscribers. It is the right approach when the problem is the shared bottleneck itself, whatever the application mix, because it moves the queue from the access equipment to the traffic optimizer, where it is actively managed.

The two are most effective when combined. On a backhaul link that saturates during peak hours, for example, Bequant can limit the most bandwidth-hungry applications and, at the same time, apply a global rate limit that guarantees a minimum bandwidth for every user. Bequant AQM additionally prioritizes interactive flows automatically, without explicit policy configuration, so gaming and video calls stay responsive under both approaches.

Yes. Traffic shaping does not need to be a permanent restriction. A permanent limit below the plan speed penalizes subscribers during off-peak hours, when the network has spare capacity, while a congestion-triggered policy only acts when a shared element is actually close to saturation.

Bequant AI congestion mitigation (ACM) works this way. It continuously monitors per-subscriber latency and losses, obtained from flows under load, and when it detects congestion symptoms it activates AQM rate limiting to keep losses and latencies low during the congestion episode. When demand falls and the congestion symptoms disappear, the limitation is released and subscribers return to their full plan speed. Because ACM is self-learning, it does not require manual fine-tuning of thresholds for each network element.

Application-based policies can follow the same logic: DPI-based limits on bandwidth-hungry applications, such as video streaming, can be focused on peak hours to moderate their demand when the network needs it, leaving them unrestricted the rest of the day. Peak hours can be identified by time of day or, more flexibly, by the overall network throughput.

Subscriber plan limits cap each subscriber individually, but they do not protect a shared element. An access point, a wireless sector or an OLT port is shared by many subscribers, and its capacity is usually lower than the sum of their plan speeds. During peak hours, every subscriber can be below their plan limit while the shared element is saturated: its queues grow, latency and losses increase, and all subscribers behind it see a degraded experience. Per-subscriber shaping alone cannot fix this, because no individual limit is being exceeded.

Topology-based shaping addresses the shared bottleneck. Subscribers are grouped by the network element that serves them (access point, sector, OLT port or backhaul link), and a rate limit matching the capacity of that element is applied to the combined traffic of the group, with scheduling that shares the available bandwidth fairly and guarantees a minimum for every subscriber. The enforcement is done in the traffic optimizer's sophisticated active queue management, instead of in the access equipment, so latency stays low and interactive flows remain prioritized.

Bequant complements this with AI congestion mitigation (ACM), which detects congestion symptoms per subscriber from flow latency and losses even when the location or capacity of the bottleneck is unknown or changing, and activates AQM rate limiting to keep losses and latencies low. Per-subscriber-group visibility shows which access points or OLT ports are suffering congestion, so the network team can decide where a topology-based policy or a capacity upgrade is needed.