Network Capacity Planning Overview
Capacity planning ensures network infrastructure can support current demand and anticipated growth without performance degradation. This overview introduces key planning concepts and frameworks.
Contents
What Is Capacity Planning
Network capacity planning is the process of assessing current and projected bandwidth, port, and compute resource requirements and ensuring that infrastructure is provisioned to meet demand with appropriate headroom. Capacity planning spans all infrastructure layers: access layer port counts, uplink bandwidth, WAN circuit capacity, internet gateway throughput, and data centre interconnect bandwidth.
Capacity planning prevents two failure modes: over-provisioning (purchasing significantly more capacity than required, producing unnecessary capital expenditure and operating cost) and under-provisioning (deploying insufficient capacity, resulting in congestion, packet loss, and degraded application performance). Both failure modes are expensive; good capacity planning minimizes both risks simultaneously.
Demand Assessment
Demand assessment establishes the baseline against which capacity decisions are made. Assessment inputs include: current traffic utilization measurements from network management systems, device population counts and types by location, planned changes in workforce size, planned application deployments and their bandwidth profiles, and video and collaboration tool usage patterns.
Traffic measurements should capture peak demand periods, not just average utilization. For most enterprise environments, peak demand occurs at defined windows: morning login periods, video conference peaks, and batch transfer windows. Capacity must be adequate for peak demand; average utilization metrics alone will understate capacity requirements for applications that users access simultaneously.
Headroom and Safety Margins
Provisioning headroom — capacity above current peak demand — serves three purposes: absorbing demand variability within planning periods, accommodating growth between planning cycles, and providing capacity during failure scenarios when redundant paths carry additional traffic. Common headroom targets for enterprise network links are 50 to 70 percent of provisioned bandwidth as the peak utilization trigger for capacity review, with upgrade decisions initiated when peak utilization consistently exceeds 70 to 80 percent.
N+1 redundancy (one spare unit beyond the count required to handle full demand) is a standard approach for critical infrastructure components including switches, uplinks, and power. N+1 provisioning ensures that a single component failure does not degrade service, but requires that each unit operate below 50 percent of its capacity under normal conditions to absorb the failed unit's load.
Planning Horizons
Network infrastructure planning typically operates across three horizons: tactical (12–18 months, addressing immediate capacity gaps and near-term growth), strategic (3–5 years, aligning infrastructure investment with technology refresh cycles and business growth plans), and long-range (5–10 years, informing building infrastructure decisions such as structured cabling capacity, conduit capacity, and electrical distribution for power-intensive deployments). Each horizon requires different levels of precision and different input sources.
Capacity Signals
Network management systems provide utilization metrics that serve as leading indicators of capacity constraints. Key metrics to monitor include: interface utilization percentages for high-capacity uplinks, queue drop rates on congested interfaces, wireless client-to-access-point ratios, and WAN circuit utilization against contracted bandwidth. Regular review of these metrics — at minimum monthly for capacity planning purposes — allows proactive identification of links approaching capacity thresholds before they affect end-user experience.