Energy saving with QoS guardrails
Cell sleep extended to the full day, with thresholds set per cell and every saving checked against real energy meters while service KPIs hold.
Results
Radio energy on a 202-site cluster, with service unchanged.
Saved per site, per day, measured over a full week.
Accessibility during the added sleep hours, same as before.
Video recap
At a glance
For
Energy, RAN optimization and sustainability teams
Replaces
A fixed night-only sleep window applied the same way to every site
Data used
PM counters, CM / configuration data, network topology and energy meter readings
Runs on
Your cloud or on-prem, every major vendor, 2G to 5G
The problem
Most operators already have cell sleep in their equipment. It usually runs in a fixed night window, around 23:00 to 06:00, the same on every site, because that is when the risk to customers looks lowest.
During the day, capacity layers stay on even when traffic drops, and they keep drawing power. Opening the window wider takes a parameter change, but nobody can say in advance which cells can sleep without hurting accessibility, drops or throughput.
So the decision stalls. As traffic moves to 5G, the LTE capacity carriers carry less each month and still cost the same to keep awake.
How it works
Measure the saving
The impact verification agent compares a full week before and after, hour against hour, on the energy meters and the service KPIs. It reports the saving per site and flags any KPI that moved.
Which sites in the southern region can sleep their capacity layer all day, and what would each one save without losing accessibility or throughput?
The sites and carriers in scope, the anchor layer that stays on, the sleep thresholds per cell and the expected saving per site, with the KPIs to watch.
Frequently asked questions
How do you know service did not suffer?
We compare the same hours a week before and a week after, weekends included, with the focus on the layer that absorbs the traffic. In the pilot, accessibility stayed at 99.96%, the drop rate at 0.04% and handover success at 99.5%.
Can every site sleep?
No. A site needs a capacity layer that can sleep and an anchor layer that stays on. The energy agent checks this site by site, so the scope is known before anything is proposed.
Does it need new hardware?
No. It uses the cell sleep feature already in your equipment. What changes is the parametrization, set per cell.
Does the platform change anything on the network by itself?
No. The agents analyse the sites, propose the parameters and measure the result. Your engineers decide, and changes go through your own systems.
Where has it been used?
With a tier-1 European operator, where 202 sites saved 4.96% of radio energy, 1.84 kWh per site per day, over a full week with service unchanged.
“O2 climbed one position in independent rankings after deploying our anomaly detection and root cause analysis engines.”
Find the hours your network can sleep.
Bring one cluster of sites. We run the agents on your data and show you which cells can sleep all day and what each site would save.
