Capacity Planning Methodology
A structured approach to data collection, analysis, modelling, and documentation for enterprise network capacity planning programs.
Contents
Data Collection
Capacity planning data is collected from network management systems, monitoring platforms, and infrastructure inventory tools. Relevant data includes interface utilization time-series at five-minute polling intervals (minimum), device population counts from network discovery tools, application traffic profiles from flow monitoring (NetFlow, sFlow, IPFIX), and planned changes from IT project portfolios.
A minimum of 90 days of historical utilization data is recommended before making capacity upgrade decisions, providing enough data to identify weekly and monthly utilization patterns. Single-week snapshots may miss utilization peaks that occur at month-end, during fiscal quarter close periods, or during annual peak business cycles.
Utilization Analysis
Utilization analysis identifies links, segments, and components operating near capacity thresholds. The primary analysis metrics are peak utilization percentages for each monitored interface over the analysis period, and the frequency and duration of periods above defined thresholds. A link that peaks at 85% of capacity for 15 minutes daily requires a different response than one that sustains 85% for three hours daily.
Separate inbound and outbound utilization analysis for WAN links, as asymmetric application profiles — common with cloud storage, video streaming, and backup traffic — may cause one direction to constrain performance while the other has significant headroom. Monitoring only aggregate or bidirectional averages can miss directional bottlenecks.
Growth Modelling
Growth modelling projects future capacity demand based on organizational growth plans and technology adoption curves. Common modelling inputs include: headcount growth by location from HR planning data, planned application deployments and their expected traffic generation, video and collaboration platform adoption (video generates 2–5 Mbps per active participant per stream), and cloud migration projects that change traffic patterns from internal to internet paths.
Simple linear growth models are appropriate for stable organizations with predictable growth. Organizations undergoing significant transformation — large-scale cloud migration, manufacturing automation, or M&A activity — require scenario-based modelling that captures step changes in traffic volume that linear extrapolation would understate.
Upgrade Thresholds
Defining capacity upgrade triggers in advance of capacity reviews — rather than responding reactively when performance problems occur — allows infrastructure teams to plan procurement and implementation in alignment with standard change management processes rather than under emergency conditions. A typical threshold framework defines: a monitoring threshold (alert when peak utilization exceeds 70%), a planning threshold (initiate upgrade planning when peak utilization exceeds 80%), and a critical threshold (escalate and expedite when peak utilization exceeds 90%).
Capacity Plan Documentation
The capacity plan is a living document updated on a defined cycle (quarterly or semi-annually for most enterprise environments). It documents current utilization status for all monitored infrastructure, growth projections for the planning horizon, upgrade recommendations with priority and timing, and assumptions underlying the projections. Capacity plans are reviewed by infrastructure management and used as input to capital expenditure planning cycles. Version-controlled documentation provides a historical record of planning decisions and the assumptions they were based on, supporting post-implementation review of forecast accuracy.