Network capacity planning is the practice of knowing how much headroom each part of your network has, forecasting when growth will use it up, and acting before it does. Most UK businesses do not do it. They discover capacity limits when something breaks: new starters arrive and there are no free switch ports, a new floor opens and the Wi-Fi struggles, video calls stutter every afternoon during a busy season, or the firewall slows to a crawl after security features are switched on. Each feels like a sudden failure. Almost none of them is.
Capacity problems are predictable because the growth that causes them is usually planned. Hiring plans, office expansions, new applications, more cameras, more video meetings and more cloud services all appear in business plans months before they reach the network. The data needed to see the limits coming — port usage, wireless client counts, firewall load, internet utilisation — is available from equipment most businesses already own. What is missing is the habit of putting the two together.
This guide starts with why capacity problems are predictable rather than sudden, then covers how to baseline current utilisation, how to forecast growth from business plans rather than guesswork, the specific components that tend to run out first — switch ports, wireless access point density, firewall throughput, internet bandwidth and some less obvious ones — and how to build a capacity review into the budget cycle instead of reacting to outages. It is written for IT managers, network administrators and business leaders responsible for networks in offices of 20 to 500 people.
What network capacity planning covers
Capacity planning is often equated with internet bandwidth, but the internet connection is only one of several components that can become a bottleneck. A useful plan covers each layer of the network.
- Switching: free ports, Power over Ethernet budgets, uplink capacity between switches and to the core.
- Wireless: access point density, client counts per access point, channel capacity and coverage in new or reconfigured spaces.
- Security and routing: firewall and router throughput with the security features actually in use, concurrent connections and VPN capacity.
- Internet and wide area connectivity: bandwidth in both directions, site-to-site links and cloud connectivity.
- Addressing and services: IP address ranges, DHCP scopes and network service limits.
- Physical infrastructure: rack space, power, UPS capacity and cooling in comms rooms, and cabling to new areas.
- Licences and subscriptions: device or user licences on cloud-managed network platforms and security services.
Each component has its own lead time to expand. Adding a switch can take days; upgrading a leased line or adding cabling to a new floor can take months. Capacity planning is largely about knowing those lead times and starting early enough.
A simple working rule: treat any component consistently above about 80% of its capacity at peak as needing action within the next planning period. Below that, there is usually room for normal fluctuations and modest growth. Above it, a hiring push, a new application or a busy season can tip it into congestion. The exact threshold varies by component, but having one turns monitoring data into decisions.
Why capacity problems are predictable rather than sudden
Capacity exhaustion feels sudden because networks often degrade gracefully until they do not. A switch with two free ports works perfectly until the third new starter. Wi-Fi with too many clients per access point feels fine at half occupancy and poor at full. A firewall running near its limit copes until a backup job, a software update and a busy afternoon coincide.
But the underlying trend is almost always visible in advance. Port usage climbs with headcount. Wireless client counts rise as people bring more devices and offices fill up. Internet utilisation grows steadily as more applications move to the cloud and video becomes standard. Firewall load increases as traffic grows and security inspection is added. None of these changes happens overnight.
The growth driving them is usually planned too. A hiring plan says the sales team will grow by fifteen people next year. A property plan says the second floor will open in the spring. A technology roadmap says the business will move its phone system to the cloud and add CCTV to the warehouse. Each has a network consequence that can be estimated months ahead.
The reason capacity problems still catch businesses out is organisational rather than technical: the people who know the business plans and the people who look after the network rarely compare notes until something has gone wrong.
Reacting to outages versus planning capacity
The comparison contrasts the two approaches most businesses fall into.
Reacting to outages
Capacity added when something breaks
Planning capacity
Headroom tracked and growth forecast
The difference in effort is modest: a few hours of data gathering and a conversation with the business each quarter or year. The difference in outcome is substantial, particularly for components with long lead times such as connectivity upgrades and cabling.
Scoring the components that run out
Different network components behave differently as they approach capacity. The score grid compares three. A high badge means a high level of that characteristic.
Switch ports give no warning — one day there are none left — but they are easy to count and quick to expand, which is why they invite quick fixes such as small unmanaged switches under desks. Wireless degrades gradually, so problems are felt before they are measured. Firewall and internet capacity have the longest lead times and the widest impact, since everyone is affected when they saturate, so they most reward early planning.
Network capacity in UK businesses — the numbers
The figures below are indicative, drawn from network reviews we carry out for UK businesses of 20 to 500 staff. They are not a formal survey.
The second figure catches many businesses out. Firewall datasheets quote headline throughput figures measured with minimal security features. Enabling intrusion prevention, malware scanning and inspection of encrypted traffic can reduce real throughput substantially, depending on the model and traffic mix. A firewall sized on its headline number can become the bottleneck as soon as its security features are used as intended.
What runs out first
The chart shows which component was the first to reach a capacity limit in the UK networks we review where capacity problems had occurred. Figures are indicative and add up to 100%.
Switch ports and PoE budgets top the list because growth in headcount and powered devices — phones, access points, cameras — consumes them directly. Wireless follows closely, especially when offices densify or new spaces open. Internet bandwidth, usually upload, comes next as video and cloud backup grow. The smaller items are worth noting precisely because they are less expected: running out of IP addresses, saturated uplinks and comms rooms without enough power or UPS capacity cause disproportionate disruption when they happen.
Baselining current utilisation
Forecasting starts with knowing where you are. A baseline records how much of each component’s capacity is in use today, at peak as well as on average, so growth can be projected from a known starting point.
Switch ports and PoE
For each switch, record total ports, ports in use, ports reserved for known purposes and spare ports, by location. Record the PoE budget and current power draw. Managed and cloud-managed switches report this directly; for others, a physical audit is needed. Note any unmanaged switches added to extend capacity, as these are a sign that planning has already fallen behind.
Wireless
Record the number of access points, their locations and the number of clients connected at peak times, along with any areas where users report poor performance. Most managed wireless platforms show client counts, channel utilisation and signal quality by access point over time. Peak hours matter more than averages.
Firewall and routing
Record peak throughput, CPU and memory utilisation, concurrent sessions and VPN connections, and compare them with the device’s rated capacity for the security features you actually have enabled, not its headline figure.
Internet and wide area links
Record peak and 95th percentile utilisation in both directions, broken down by time of day and day of week, and note recurring peaks such as backups or software updates. Our guide to bandwidth planning for growing businesses covers sizing the internet connection in detail.
Addressing, uplinks and physical infrastructure
Record DHCP scope usage by network segment, uplink utilisation between switches, rack space, power and UPS load in each comms room, and licence counts on network platforms.
Use monitoring, not snapshots
A single snapshot can mislead, because utilisation varies through the day, week and year. Network monitoring that records trends over months gives a far more reliable baseline and shows the rate of growth directly. Our guide to proactive network monitoring explains what to collect.
Baseline data collected in UK networks
The progress bars show how often each type of capacity data is collected and retained over time in the UK networks we review. Figures are indicative.
Capacity data collected and trended
Internet utilisation and wireless client counts are the most commonly tracked, often because the platforms show them by default. PoE budgets, DHCP scopes, uplinks and UPS load — the components that cause the least expected failures — are tracked far less often. And very few businesses take the final step of comparing any of this with business plans, which is what turns data into a forecast.
Forecasting growth from business plans rather than guesswork
Infrastructure growth forecasting is most reliable when it starts from what the business intends to do, translated into network terms, rather than from a general assumption that “usage grows by some percentage a year”.
Gather the plans
Collect the business plans that drive network demand: hiring plans by team and location, office changes such as new floors, moves or consolidations, new applications or systems, changes in working patterns such as more video or more hybrid working, new devices such as CCTV, sensors, digital signage or production equipment, and seasonal patterns.
Translate plans into network demand
Convert each plan into its effect on each component. Fifteen new desk-based staff might mean fifteen switch ports, fifteen or more wireless clients, perhaps fifteen phones drawing PoE, more video calls and more internet bandwidth. A warehouse CCTV project might mean twenty PoE ports with significant power draw, plus bandwidth for recording and remote viewing. A move to a cloud phone system might mean more internet bandwidth and stricter quality requirements. Simple rules of thumb per person, per device and per application, refined from your own baseline, make this quick.
Add organic growth
On top of planned changes, allow for organic growth: more devices per person, larger files, more cloud services, higher video quality. Your trend data shows the historic rate; use it rather than guessing.
Project against capacity
For each component, project utilisation forward over the next twelve to twenty-four months and identify when it will cross your threshold. Compare that date with the lead time to expand. If the threshold will be crossed in six months and the upgrade takes four, action is needed now.
Allow for uncertainty
Business plans change. Build forecasts with a realistic and a higher-growth scenario, and favour expansions that can be scaled in steps rather than single large commitments.
Rules of thumb for translating plans into numbers
Forecasting does not need complex models. A small set of rules of thumb, refined from your own baseline, turns business plans into network demand quickly. The figures below are illustrative starting points; replace them with values from your own monitoring as soon as you have them.
- Desk-based staff: typically one or two switch ports per desk (computer and phone, or docking station), two to three wireless clients per person, and a share of internet bandwidth informed by your current per-person peak.
- Hybrid staff: plan wireless and bandwidth for the proportion expected on the busiest day, not the total headcount.
- Desk phones: one port and a PoE draw each, plus voice bandwidth and quality-of-service requirements.
- Wireless access points: one PoE port each, with newer models drawing more power; plan density from peak clients and the applications in use.
- Cameras: one PoE port each, with higher draw for models with motors or heaters, plus recording and remote viewing bandwidth.
- Meeting rooms: ports for room systems and displays, wireless capacity for attendees, and upload bandwidth for simultaneous video calls.
- New cloud applications: estimate from vendor guidance and your experience with similar services, and watch the first months closely.
Multiply each rule by the planned change, add up the effect on each component, and compare the total with remaining headroom. The result does not need to be precise to be useful; it needs to show which components will run short, and roughly when.
Seasonal and event-driven peaks
Some capacity problems are tied not to steady growth but to predictable peaks. Retailers see traffic surge before Christmas; accountancy firms around tax deadlines; education businesses at the start of term; recruitment firms in January. Events such as company-wide meetings, product launches or training days can also bring unusual numbers of people and devices into the office at once.
Look at last year’s utilisation data for these periods and plan capacity for them specifically. Sometimes a temporary measure is enough — additional bandwidth for a quarter, portable wireless for an event, scheduling backups and updates outside peak times. Sometimes the peak defines the capacity the business needs all year. Either way, a known peak should never come as a surprise.
An annual capacity review cycle
The timeline sets out a practical annual cycle that ties capacity planning to the budget process, so upgrades are planned and funded rather than reactive.
The cycle does not need to be elaborate. For a business of 50 to 200 staff, the annual review might take a day or two, with quarterly check-ins of an hour. Its value lies in timing: identifying needs before the budget is set, so they are funded, and ordering long-lead items early enough to avoid emergencies. Changes should then go through normal change control; our guide to network change management covers making them safely.
Capacity planning maturity — where most UK businesses sit
Combining baseline data, trend monitoring, forecasting from business plans, understanding of lead times and integration with the budget cycle gives an indication of capacity planning maturity. The gauge reflects a typical UK business network of 20 to 500 users.
A score in the high twenties typically reflects a network where some utilisation data exists, usually internet bandwidth and wireless clients, but nobody reviews it against business plans, PoE budgets and firewall load are unknown, and upgrades happen when something breaks. Capacity is managed by complaint.
Collecting baselines for each component, holding one annual review aligned with the budget and a few quarterly check-ins typically lifts maturity into the sixties within a year. The improvement shows less in metrics than in what stops happening: emergency purchases, unmanaged switches under desks and the busy-season slowdown everyone had learned to expect.
Planned expansion versus emergency upgrades
The table compares indicative costs of planned capacity expansion with the typical cost of emergency upgrades for common components in a UK office of 50 to 200 staff. Figures are broad ranges and depend on equipment choices, location and supplier.
| Component | Planned expansion | Emergency response | Hidden cost of reacting |
|---|---|---|---|
| Switch ports and PoE | £800 to £3,000 per managed switch, installed in a planned window | Unmanaged switches under desks, or rushed purchases at premium | Unmanaged devices weaken security and visibility |
| Wireless access points | £300 to £900 per access point plus cabling, with a survey | Access points added where convenient, without design | Interference and coverage gaps persist |
| Firewall | £2,000 to £10,000 or more, aligned with refresh and licensing | Security features disabled to recover throughput | Reduced protection until replacement |
| Internet bandwidth | Planned upgrade or second circuit within normal lead times | Temporary connections at premium, or months of congestion | Lost productivity and poor call quality |
| Cabling to new areas | Installed during fit-out at normal rates | Retrofitted after occupation, with disruption | Higher cost and visible trunking |
The pattern repeats for every component: planned expansion costs roughly what it costs, while reactive expansion costs that plus a premium, plus a period of degraded service, plus compromises such as unmanaged switches or disabled security features that create longer-term problems. The largest differences arise where lead times are long, especially connectivity and cabling.
Presenting the capacity plan to leadership
Capacity upgrades compete for budget with everything else, and requests framed in technical terms tend to lose. Frame them in business terms instead.
Link each recommendation to a business plan: “The sales team expansion in the spring needs twelve more switch ports and more wireless capacity on the second floor.” State the consequence of not acting: “Without it, new starters will not have network connections, and afternoon video calls will continue to degrade.” Give the cost, the lead time and the latest date to decide. Where possible, show how combining capacity with a planned hardware refresh reduces cost.
Keep it short. One page listing each component, its headroom, the forecast date it will run short, the recommended action and the cost is usually more persuasive than a long technical report. Revisit it quarterly so leadership sees forecasts being tracked and updated, which builds confidence in future requests.
How often components are near capacity without anyone knowing
Across the UK networks we review, the proportion in which at least one key component is consistently above 80% of capacity at peak, without the business being aware of it, is substantial.
The figure is indicative rather than a formal survey. The most common culprits are PoE budgets on switches feeding access points and cameras, firewalls running near their limit with security features enabled, and upload bandwidth during working hours. In each case the network is still working, which is why nobody has noticed, but it has little headroom left for growth or unusual demand.
The components that run out first, in detail
Switch port exhaustion
Switch port exhaustion is the most common capacity problem and the most avoidable. Ports are consumed by desks, phones, printers, access points, cameras, door controllers and an ever-growing list of devices. Keep a margin of spare ports in each location, typically ten to twenty per cent, and plan additional switches before a hiring push or office change. Avoid the quick fix of unmanaged switches under desks, which are invisible to monitoring, cannot be secured properly and can cause network-wide problems.
Power over Ethernet budgets
Each PoE switch has a total power budget shared across its ports. Newer wireless access points, pan-tilt-zoom cameras and other powered devices draw more power than older ones, so a switch can run out of power before it runs out of ports. Check PoE budgets when adding powered devices or upgrading access points, and consider this when choosing switches.
Wireless access point density
Wireless capacity depends on the number of clients per access point, the applications they use and how the radio spectrum is shared. Densifying an office, opening collaboration spaces or adding meeting rooms with video can overload access points designed for a lighter load. Plan wireless for peak occupancy and the applications in use, ideally with a survey when spaces change, and remember that adding access points without design can make performance worse through interference.
Firewall throughput
Firewalls must handle total traffic with all required security features enabled. Growth in traffic, more encrypted traffic to inspect, additional VPN users and new security services all increase load. Size firewalls on realistic throughput with the features you use, with headroom for growth over their expected life, and monitor CPU and session load.
Internet bandwidth
Internet demand grows with cloud services, video and backups, and upload is often the constraint. Lead times for upgrades can be long, so forecast well ahead and consider a second connection for resilience as well as capacity.
IP addressing and DHCP
Address ranges sized for an office of fifty can run out when guest Wi-Fi, phones, IoT devices and staff devices multiply. Exhausted DHCP scopes cause devices to fail to connect, often intermittently. Review address plans when segments grow, and size new ranges generously.
Uplinks, power and licences
Uplinks between access switches and the core can saturate as more devices and faster wireless push more traffic. Comms rooms can run out of rack space, power or UPS runtime as equipment is added. Cloud-managed platforms and security services may be licensed per device or user, and running out of licences can stop new equipment working. Include each in the baseline and review.
Hardware refresh cycles interact with all of these: replacing ageing switches, access points or firewalls is the natural moment to add capacity. Our guide to network hardware lifecycle and replacement covers planning refreshes.
The 12-point network capacity planning checklist
The checklist covers the essentials for keeping ahead of network capacity.
- Inventory every capacity-limited component. Switches, PoE budgets, access points, firewalls, circuits, address ranges, uplinks, comms rooms and licences.
- Baseline utilisation at peak. Not just averages, and upload separately from download.
- Monitor trends continuously. With history kept long enough to show growth rates.
- Set thresholds. Around 80% at peak as a prompt for action, adjusted per component.
- Know each component’s lead time. Days for a switch, months for a circuit or cabling.
- Gather business plans. Hiring, property, applications, devices and seasonal patterns.
- Translate plans into network demand. Ports, power, clients, throughput and bandwidth per change.
- Forecast twelve to twenty-four months ahead. With realistic and higher-growth scenarios.
- Align upgrades with refresh cycles. Add capacity when replacing ageing equipment.
- Review annually before the budget. So upgrades are funded, not emergencies.
- Check quarterly against plan. Adjust timings as business plans change.
- Avoid bodges. No unmanaged switches or disabled security features as capacity fixes.
The most effective capacity plans fit on a single page: each component, current headroom, forecast date to reach the threshold, lead time, recommended action and cost. Presented before the budget is set, a page like this turns network upgrades from unexpected requests into planned investments that leadership can weigh against business growth.
Case study: the hiring push that the network nearly missed
A UK recruitment business with around 140 staff across two floors planned to grow its consultant headcount by about forty people over the coming year, opening a third floor in the same building in the spring. The IT team learned about the plan informally, two months before the new floor was due to open.
A capacity review found several issues. The existing floors had fewer than ten spare switch ports between them, and several unmanaged switches had already been added under desks. PoE budgets on the switches feeding phones and access points were above 85%. Wireless client counts at peak were already high on the busier floor, with complaints about video calls in the afternoons. The firewall, sized on its headline figure, was running at around 75% CPU at peak with intrusion prevention enabled. Internet upload was regularly saturated during the late afternoon when consultants held video interviews. The third floor had no network cabling at all.
With two months to go, the business prioritised: cabling and switches for the new floor were ordered immediately, a wireless survey informed access point placement, the unmanaged switches were replaced with managed ones, and PoE-capable switches with larger budgets went into the existing floors. The internet upgrade could not be delivered in time, so a second connection was added on a shorter lead time and traffic balanced across both. The firewall replacement was brought forward from the following year.
The new floor opened on time, but at higher cost than a planned approach would have needed. The business then introduced an annual capacity review aligned with its budget cycle, with HR and property plans shared with IT each quarter. The following year’s growth was absorbed with planned, budgeted upgrades and no emergencies.
“Every one of those problems was visible in the data months earlier. We just never looked at the network and the hiring plan in the same meeting. Now we do, every quarter, and the network is ready before the people arrive.” — Operations Director, UK recruitment business
The lesson is not that the team did anything wrong. It is that capacity planning depends on connecting business plans to network data, and that connection needs to be built into how the business plans its year.
Common network capacity planning mistakes
The mistakes below account for most of the capacity problems we see.
- Planning only internet bandwidth. Ports, PoE, wireless and firewalls often run out first.
- Using averages instead of peaks. Congestion happens at peak times.
- Sizing firewalls on headline throughput. Security features reduce real capacity substantially.
- Ignoring upload. Video and cloud backups make upload the common constraint.
- Forgetting PoE budgets. Switches can run out of power before ports.
- Adding access points without design. More radios can mean more interference.
- Unmanaged switches as a fix. Invisible, insecure and a source of network-wide issues.
- Not knowing lead times. Circuits and cabling take months.
- Not talking to the business. Hiring and property plans arrive too late.
- Leaving capacity out of the budget. Every upgrade becomes an unplanned request.
When a firewall becomes a bottleneck, a tempting quick fix is to disable security features such as intrusion prevention or inspection of encrypted traffic to recover throughput. This leaves the business less protected precisely when it is busiest. Treat a firewall near capacity as a planned upgrade, not a reason to switch protection off.
How cloud and hybrid working change capacity needs
Network scalability planning looks different now than it did when most applications ran on servers in the office. Several shifts change where capacity is needed.
More internet, less internal traffic. With email, files and applications in the cloud, traffic that once stayed inside the office now crosses the internet connection. Internet bandwidth, especially upload, and firewall throughput become more important, while internal server links matter less.
Video everywhere. Video calls are now part of almost every working day, and hybrid meetings mean many people in the office join calls simultaneously. This drives demand for upload bandwidth, wireless capacity in meeting and collaboration spaces, and quality of service.
Peakier occupancy. Hybrid offices are often quiet on some days and full on others. Capacity must be planned for the busiest days, not the average, and those peaks can be higher than in a traditional office as teams coordinate their days in.
More devices per person. Laptops, phones, tablets and wearables all connect to Wi-Fi, increasing client counts and address usage even where headcount is flat.
More connected building systems. Cameras, access control, sensors, displays and smart building systems consume ports, power and addresses, and often belong on separate network segments.
Capacity planning for new offices and fit-outs
The cheapest time to add capacity is before a space is occupied. When opening a new floor, fitting out a new office or reconfiguring a space, plan the network for at least the life of the lease, not just day one.
Install cabling generously during fit-out, since adding it afterwards is disruptive and expensive. Size comms cabinets, power and UPS for growth. Design wireless for the planned layout and peak occupancy, with a survey where possible. Choose switches with spare ports and PoE budget for the devices you expect to add. Order connectivity early, allowing for long lead times. Our guide to planning network infrastructure for a growing business covers the design side in more depth.
Capacity planning across multiple sites
Businesses with several offices, branches or warehouses need capacity planning at each site and across the connections between them. Growth rarely happens evenly: one site may be expanding while another shrinks. Baseline each site separately, and include site-to-site links, cloud connectivity and any central services that branches depend on.
Standardising equipment and designs across sites makes planning easier, because the same rules of thumb and lead times apply everywhere. Cloud-managed network platforms help by showing utilisation across all sites in one place. When sites change role — a branch becoming a regional hub, or a warehouse adding automation — treat it as a capacity event and review that site specifically.
Who should own capacity planning
Capacity planning falls between roles: IT has the data, the business has the plans, finance has the budget. Without a named owner, it does not happen. In smaller businesses, the IT manager or managed service provider is the natural owner, with a commitment from leadership to share plans that affect headcount, space and systems. In larger organisations, capacity planning may sit with infrastructure or architecture leads, feeding into the annual IT plan.
Whoever owns it, two inputs are essential: regular access to business plans, ideally through a standing agenda item in quarterly planning, and monitoring data that is kept and trended rather than glanced at during incidents. With both, capacity planning becomes a modest, routine task rather than a periodic crisis.
At a glance: network capacity planning
| Component | What to measure | Typical lead time to expand | Common warning sign |
|---|---|---|---|
| Switch ports | Ports used and spare by location | Days to weeks | Unmanaged switches appearing |
| PoE budget | Power drawn against switch budget | Days to weeks | Devices failing to power up |
| Wireless | Clients per access point, channel use at peak | Weeks, with survey | Afternoon slowdowns, poor calls |
| Firewall | Throughput, CPU, sessions with features enabled | Weeks | High CPU at peak, slow browsing |
| Internet bandwidth | Peak and 95th percentile, upload and download | Weeks to months | Video quality drops at busy times |
| IP addressing | DHCP scope usage by segment | Days, with planning | Devices intermittently failing to connect |
| Uplinks | Utilisation between switches and core | Days to weeks | Slowdowns affecting whole floors |
| Comms room | Rack space, power, UPS load and runtime | Weeks to months | Shortened UPS runtime, full racks |
| Licences | Device and user counts against entitlements | Days | New equipment will not register |
| Cabling | Outlets and routes in new or changed areas | Weeks to months | Requests for desks without outlets |
| Overall process | Plans versus utilisation trends | Annual and quarterly | Upgrades always urgent and unbudgeted |
How Cloudswitched approaches network capacity planning
We build capacity planning into ongoing network administration rather than treating it as a one-off exercise. That starts with a baseline across every capacity-limited component — switch ports and PoE budgets, wireless density, firewall load with security features enabled, internet bandwidth in both directions, address ranges, uplinks, comms room power and licences — and continuous monitoring that keeps the history needed to see trends.
Each year, ahead of your budget, we gather your hiring, property, application and device plans, translate them into network demand, and forecast when each component will need attention against its lead time. The result is a short, prioritised capacity plan with costs and timings, aligned with hardware refresh cycles so capacity is added at the natural moment. Quarterly check-ins keep the plan aligned with changes in the business, and upgrades are delivered through normal change control before demand arrives.
No forecast is perfect, but a network managed this way rarely surprises anyone. Growth is absorbed with planned, budgeted upgrades rather than emergency purchases and workarounds.
Want your network ready before growth arrives?
Talk to us about a network capacity baseline, forecasting from your business plans and an annual capacity review built into your budget cycle.
Talk to a Network Administration SpecialistFrequently Asked Questions
What is network capacity planning?
The practice of measuring how much of each network component’s capacity is in use, forecasting how business growth will consume it, and expanding capacity before limits are reached, rather than after something breaks.
Which parts of the network run out of capacity first?
Most commonly switch ports and PoE budgets, then wireless capacity, internet bandwidth (often upload) and firewall throughput. IP address ranges, uplinks and comms room power are less common but disruptive.
How do we forecast network growth?
Start from business plans — hiring, office changes, new applications and devices — and translate each into ports, power, wireless clients, throughput and bandwidth. Add organic growth based on your trend data, and project forward against capacity.
What utilisation level should trigger an upgrade?
A common working threshold is around 80% of capacity at peak, adjusted for each component. Consider the lead time to expand: act early enough that upgrades complete before the threshold is crossed.
Why does our firewall slow down when security features are enabled?
Inspection features such as intrusion prevention and examining encrypted traffic require significant processing. Real throughput with these enabled is often well below headline datasheet figures, so size firewalls on realistic figures.
What is switch port exhaustion?
Running out of free ports on network switches, usually as headcount and connected devices grow. It is common and easy to avoid by tracking spare ports and adding switches ahead of need.
Are unmanaged switches a problem?
Using them to work around port shortages is. They are invisible to monitoring, cannot be secured or configured properly, and can cause network-wide issues. Replace them with managed switches as part of planned expansion.
How does hybrid working affect network capacity?
It shifts demand to internet bandwidth and upload, increases video traffic and wireless load in meeting spaces, and creates peakier occupancy. Plan for the busiest days rather than averages.
How often should we review network capacity?
An annual review aligned with the budget cycle, with quarterly check-ins against business plans and continuous monitoring with threshold alerts, works well for most businesses.
Do we need special tools for capacity planning?
Not usually. Managed switches, wireless platforms, firewalls and network monitoring tools most businesses already have provide the data. The key is retaining history and reviewing it against business plans.
What is a PoE budget?
The total power a switch can supply to connected devices such as phones, access points and cameras. Adding or upgrading powered devices can exhaust it before the switch runs out of ports.
How much spare capacity should we keep?
As a guide, ten to twenty per cent spare switch ports per location, wireless designed for peak occupancy, and firewalls and internet connections with headroom for at least the next year’s forecast growth.
Related reading
More guidance for UK businesses from the Cloudswitched blog:
- Office downsizing and consolidation: moving to a smaller office without losing IT capability
- Private connectivity to Azure and AWS: ExpressRoute and Direct Connect vs VPN
- Backup encryption: protecting backup data without locking yourself out
- Penetration testing for web applications: what an OWASP-based test covers
- Microsoft 365 Copilot adoption: why most deployments fail to change how people work
No more surprise bottlenecks
Cloudswitched baselines, forecasts and plans network capacity for UK businesses, so ports, Wi-Fi, firewalls and bandwidth are ready before growth needs them.
Talk to a Network Administration Specialist
Cloudswitched
London-based managed IT services provider offering support, cloud solutions and cybersecurity for SMEs.
