A network infrastructure refresh is the planned replacement of aging switches, routers, firewalls, wireless and cabling with current equipment, sequenced so the business stays connected throughout. It runs in five phases: assessment, design, physical preparation, implementation and decommissioning of the old gear. Most failed refreshes skip or rush the first and third phases.
When does a network need a refresh?
A network needs a refresh when the cost of keeping old equipment running, in staff time, risk or lost performance, exceeds the cost of replacing it. These indicators usually show up together:
- Performance degradation: increasing latency, packet loss or bandwidth constraints affecting business applications
- Security vulnerabilities: equipment running firmware that no longer receives security updates
- Operational overhead: growing time spent on maintenance, workarounds and troubleshooting
- Scalability limits: infrastructure that cannot accommodate business growth or new technologies
- Compliance gaps: equipment that doesn't meet current regulatory requirements
Firmware support status is the clearest trigger. Once a platform stops receiving security patches, every month it stays in service adds risk you cannot patch away. Check vendor end-of-support dates for every model on the network, not just the core, and build the refresh schedule around the earliest ones.
What should you assess before buying new equipment?
Assess what you have, what depends on it and how it performs today. Automated discovery tools help, but manual verification catches what automation misses.
Network mapping. Document every device, connection and configuration, including:
- Physical topology and cable paths
- Logical network segmentation
- IP addressing schemes
- Traffic patterns and bandwidth utilization
- Quality of service configurations
Application dependency mapping. Modern networks support complex application ecosystems. Map which applications depend on which network paths, identify latency-sensitive traffic and document failover requirements. Mapping application dependencies before migration explains the method.
Performance baselining. Measure throughput, latency, jitter and error rates across different network segments now, while the old network is in service. That baseline is the only way to prove later that the new network performs better rather than merely differently.
How much capacity should the new network be designed for?
Design for the bandwidth you expect to need over the next five to seven years, not for today's peak. Consider emerging demands such as IoT, increased video conferencing and cloud application growth. Building in headroom avoids premature obsolescence.
Then evaluate vendors and platforms against your specific requirements:
- Throughput and port density needs
- Management and automation capabilities
- Security feature integration
- Vendor support and longevity
- Total cost of ownership
Security features deserve a written policy before you compare products. NIST's guidance on firewall policy (SP 800-41) is a practical reference for deciding what the new perimeter and internal segmentation must enforce, which in turn tells you which platforms qualify. Ask for published end-of-support dates on every shortlisted model, so the new network does not start life halfway to its next refresh.
Parallel, phased or flash cut: which migration strategy fits?
Pick the strategy by how much downtime the business tolerates and how much space and power you have to run old and new side by side.
| Strategy | How it works | Downtime | Main requirement | Best fit |
|---|---|---|---|---|
| Parallel deployment | New network runs alongside the old until validated | Minimal | Rack space, power and ports for both | Core and data center networks |
| Phased migration | Segments upgraded one at a time | Short, per segment | Clear segment boundaries | Campus and multi-floor sites |
| Flash cut | Complete replacement in one maintenance window | One planned outage | Thorough staging and a rehearsed rollback | Small sites, branch offices |
Whichever you choose, identify what could go wrong and plan countermeasures: rollback procedures, saved copies of current configurations held by your network team, and contingency time in the schedule. A flash cut without a tested rollback is a bet, not a plan.
Will your cabling, power and cooling support the new equipment?
Often not without changes, which is why physical preparation is its own phase.
Cabling. Evaluate whether existing cabling supports new equipment speeds. Category 5e may suffice for gigabit, but higher speeds require Category 6A or fiber. The IEEE 802.3 Ethernet working group publishes the standards that define which media each speed supports. Plan cable runs for new equipment locations, and document them; see cable documentation that survives the move.
Power and cooling. Modern switches and routers may have different power requirements and heat output. Verify that racks have adequate power circuits and cooling capacity, and check intake temperatures against the ASHRAE TC 9.9 recommended envelope for IT equipment.
Rack space. Document rack elevations for new equipment. Plan cable management to maintain airflow and accessibility, and leave room for expansion. If the closets themselves are moving, structured cabling and cross-connects for a new build covers the pathway work.
How should new network equipment be staged and deployed?
Stage and test first, then deploy from the least critical segment toward the core.
Staging and testing. Your network team configures new equipment in a lab environment before deployment, tests configurations, verifies performance and validates failover mechanisms in a controlled setting.
Deployment sequencing. Follow your migration strategy, typically starting with less critical segments to validate the process before tackling core infrastructure.
Physical installation. Racking, power, patching and labeling follow the rack elevations and port maps from the planning phase. This physical work can be done by your own staff or an installation crew, while configuration stays with your network team.
Documentation updates. Update network documentation as changes occur. Outdated documentation creates operational risk and troubleshooting delays.
Validation and optimization. After deployment, your network team confirms performance meets or exceeds the baseline and fine-tunes configurations against actual traffic patterns and application behavior.
What happens to the old network equipment?
Old equipment should leave the building clean, documented and accounted for.
Configuration removal. Your network team clears all configurations from decommissioned equipment to prevent security exposure and intellectual property leakage. Saved configurations can reveal addressing, credentials and firewall rules, so treat a retired router like a retired drive.
Asset disposition. Follow your organization's disposal policies. Options include certified recycling, resale through proper channels, or redeployment to less critical roles. Update your asset records as each unit leaves, and keep certificates for anything recycled or destroyed. IT asset disposition vendor selection covers what to ask a disposition provider.
Documentation archival. Retain old network documentation for historical reference and compliance before removing it from active systems.
Schedule removal promptly. Old equipment left racked "just in case" consumes power and space, confuses the next audit and tends to stay connected long after anyone remembers why.
What separates successful network refreshes from troubled ones?
Successful refreshes share five characteristics, and none of them is technical:
- Executive sponsorship: leadership understands and supports the project scope
- Cross-functional involvement: application owners, security teams and business stakeholders take part from the start
- Realistic timelines: adequate time for testing and validation
- Communication plans: affected users know about changes and potential impacts in advance
- Professional assistance: experienced partners for complex migrations, physical installation or relocation
Troubled refreshes fail in reverse: the core cut over first, cabling surprises on install night, users learning of an outage as it happens. For refreshes that coincide with moving the network to a new building, see the network infrastructure relocation guide.
A network infrastructure refresh is a significant undertaking, but with careful planning and execution it delivers better performance, stronger security and a foundation for future growth.