/// TELECOM · CAPACITY PLANNINGNetwork Capacity Planning & Forecasting
Traffic forecasts by cell, node and route, with the uncertainty shown and your upgrade lead times built in — so capacity arrives before congestion does.
Message on WhatsApp
TRUSTED BY TEAMS
Network capacity planning decides where and when a network needs more capacity before demand outgrows it. Getting it wrong is expensive either way: build too early and capital sits idle; build too late and customers suffer congestion, complain and leave.
Sodio builds network traffic forecasting and capacity planning for regional ISPs, fibre operators, MVNOs and CPaaS providers — forecasts for every node, shown with their uncertainty, timed against how long each upgrade really takes to deliver.
/// IS IT WORTH BUILDING YET?
When Forecasting Is Worth Building
Tier-one operators run capacity planning inside large vendor suites, and this is not a replacement for those. For smaller operators, the question is whether a forecast would change what gets built, and when.
Worth building when
- Thousands of nodes, links or cells need forecasting
- Demand has strong daily, seasonal or event-driven peaks
- Upgrades take months, so late warnings are costly
- Capital is tight and must go where congestion risk is highest
A spreadsheet will do, when
- A small network where planners know every link
- Traffic grows slowly and steadily
- Capacity can be added quickly and cheaply
- There is spare capacity everywhere for the foreseeable future
/// WHAT IT COVERS
What Network Capacity Planning Covers
Traffic forecasting by node
Uncertainty ranges
Seasonality and events
Growth drivers
Threshold alerts
Lead-time planning
What-if scenarios
Upgrade prioritisation
Wholesale commitment forecasting
Forecast accuracy tracking
/// THE DATA IT USES
The Data Behind the Forecast
///LEAD TIME
Forecasting That Knows Its Lead Time
A forecast is only useful if it arrives before the decision it informs. If a fibre upgrade takes six months to deliver, a warning that congestion will start next month is too late to act on. So every forecast is read against the lead time of the upgrade it would trigger, and the alert fires when the work needs to start, not when the link is already full.
Forecasts are shown as ranges rather than single figures. A planner deciding whether to spend capital needs to know not just the likely demand, but how wrong the forecast might be — and which way.
Deliberate capital trade-offs
Both mistakes cost money. Build too early and capital sits idle; build too late and customers leave. The point of forecasting is to make that trade-off deliberately.
Both mistakes cost money. Build too early and capital sits idle; build too late and customers leave. The point of forecasting is to make that trade-off deliberately.
/// FAQ
Frequently Asked Questions
It is deciding where and when a network needs more capacity — more bandwidth on a link, another carrier on a cell, a new fibre port or a larger wholesale commitment — before demand outgrows it. It depends on a forecast of traffic at each point in the network, and on knowing how long each kind of upgrade takes to deliver.
It depends on how much history you have and how stable your network is. Short-range forecasts, hours to days ahead, support operations. Planning forecasts run months ahead to match upgrade lead times. Accuracy falls the further out you look, which is why every forecast comes with a range that widens with the horizon.
For a small, stable network, trend lines may be enough, and we will say so. Machine learning earns its place when there are thousands of nodes to forecast, strong daily and seasonal patterns, one-off events, and growth that does not follow a straight line.
Utilisation history for each node or link — ideally at least a year, so seasonal patterns are captured — plus network topology and subscriber or customer growth. Most of it already sits in your network monitoring and OSS/BSS systems.
Yes. The points being forecast differ — fibre ports, routers, cells, backhaul links — but the method is the same. For MVNOs, which do not own the radio network, the same forecasting applies to wholesale capacity commitments with the host operator.
It can forecast 5G traffic and show where capacity or densification will be needed. Radio frequency planning — modelling coverage and signal propagation for new sites — is a separate discipline handled by dedicated RF planning tools, and we integrate with them rather than replace them.
/// RELATED SERVICES
Related Industries and Services
Routes where scope overlaps with network capacity planning.
/// GET STARTED
Start With the Links That Congest First
Share a year of utilisation data for part of your network. We will assess, free, how accurately its demand can be forecast, and how much earlier your team would have known where to build.