Re-cabling a building is expensive, disruptive and hard to reverse. It is also the first thing people reach for when a network feels slow, which is why so much of it gets done for the wrong reasons. Here are the triggers that actually justify it, and the ones that do not.
Figure 1: Structured cable trays and vertical risers upgraded to 10G capability.Quick answer
Trigger | Does it justify a re-cable? |
|---|
10GBASE-T required to the desk | Yes — and to Cat6A, not Cat6 |
|---|
Access points moving to 2.5G or 5G | At the far runs, often yes |
|---|
PoE devices in a large bundle | Depends on the bundle and the heat, not on the category alone |
|---|
The ceiling or walls are already open | Yes. Do it while the access is free |
|---|
Everything runs at 1G and will continue to | No |
|---|
Links failing certification | Re-terminate first, then decide |
|---|
A new tenant with a written bandwidth requirement | Yes, if the requirement exceeds Cat5e |
|---|
Start with what the cable can already do
Cat5e is not a slow cable. It carries 1G over the full 100-meter channel and it carries 2.5GBASE-T over the same distance. For a site where the client devices are on Gigabit and the access points are on Gigabit, an existing Cat5e installation is not the constraint — and replacing it changes nothing that anybody will measure.
The upgrade conversation only becomes real when something in the requirement moves past what 100 MHz can hold. That is either a rate above 2.5G at the access point, or 10GBASE-T anywhere beyond 2.5 meters.
Four triggers that justify the work
1. A genuine 10-Gigabit requirement. If the design needs 10GBASE-T to the desk, Cat5e and Cat6 both fail, and the answer is Cat6A. That is a re-cable, and the specification should be Cat6A rather than the next rung up.
2. Access points outgrowing their uplink. Access points now commonly present 2.5G or 5G ports. Cat5e handles 2.5G at full distance and not 5G. If the wireless design calls for 5G uplinks, the long access point runs need Cat6, and the short ones may not.
3. An open ceiling or an open wall. This is the strongest practical trigger. If a fit-out is happening anyway, or a ceiling is down for another trade, the marginal cost of pulling Cat6 is small. Waiting until the requirement appears means paying the full disruption cost later.
4. Airflow, density and heat in the rack or pathway. PoE in a large bundle raises the temperature of the cables, and older thinner cable has less margin for it. Where a bundle is being reworked for PoE load, upgrading the pathway at the same time is reasonable.
Three triggers that do not
A slow-feeling network with 1G ports on both ends. The cable is not the problem. Look at the switch uplinks, the WAN, and whether the patch leads are the wrong grade.
A single failing link. Re-terminate it. Most link failures are terminations rather than cable, and a failed termination sits in exactly the same place in the channel whether the cable behind it is Cat5e or Cat6.
"We might need 10G one day." That is a reason to specify Cat6 or Cat6A for new work. It is not a reason to re-cable a functioning Cat5e building, because by the time the need arrives, the products on offer will have moved.
Do it in phases if it has to be done
A building-wide re-cable is rarely the only option. Three cheaper shapes exist.
Upgrade the runs that will need it. The long runs to the far corners and the access point positions are usually the ones that hit a limit first. Short desk runs inside the 55-meter 10G window may be fine on Cat6 or may not need changing at all if 1G is their fate.
Upgrade the pathway, not every outlet. Pull spare capacity through the main routes during the work. Terminating later is much cheaper than pulling later.
Upgrade the components you touch. Where a panel is being replaced or an outlet re-terminated, doing it with Cat6 components at least stops the channel getting worse. Note that a Cat6 component on a Cat5e cable still caps the channel at Cat5e — the improvement is durability and consistency, not performance.
What the work actually involves
A re-cable is not only cable. It is the cable, the outlets, the patch panel, the labels, the certification and the downtime. Several of those cost more than the cable by a wide margin, and the certification is the one most often left out of the budget until the end.
The certification step is also what turns the work into a documented asset. A channel-by-channel report at handover is what lets somebody answer "which outlets support 10G" two years later without opening a ceiling.
The cost of not upgrading
There is a real one, and it is worth naming. A building that cannot carry 5G to its access points caps the wireless performance of everyone in it, regardless of the access points installed. A building that cannot carry 10G to a workstation caps the workstation. And the cost of the re-cable does not fall over time — the disruption is the same, and it is usually higher when the building is occupied.
The judgment is about timing rather than direction. Upgrading during planned work costs a fraction of upgrading in an emergency.
Field scenarios
A school with Cat5e, 1G switches and 1G access points, planning a wireless refresh. First question: will the new access points want 2.5G uplinks? If yes, the long runs need attention, and the short ones may not. If no, the cabling is fine and the money belongs in the wireless.
An office tenant whose lease includes a fit-out. Specify Cat6 for the horizontals while the ceiling is open. It is the cheapest moment the decision will ever have.
A plant with Cat5e in a warm ceiling void and PoE cameras on long runs. Measure first. Temperature and bundle effects can make a cable that should work fail, and the answer may be re-routing rather than re-cabling.
FAQ
Is Cat6 better than Cat5e?
Yes, in bandwidth and in the rates it supports at full distance: 5G where Cat5e stops at 2.5G, and 10G over short runs. Whether that matters depends entirely on what the network will be asked to carry.
Is Cat5e better than Cat6?
Only on price, and only where 1G or 2.5G is the ceiling for the life of the installation.
Can I upgrade only some runs?
Yes, and it is often the right approach. Document which channels are which grade at the panel.
Do I need to replace the patch panel?
If the panel is Cat5e, it caps the channel at Cat5e whatever cable is behind it. Replace it if the channel has to perform above Cat5e.
Will a re-cable make the internet faster?
Only if the local network was the bottleneck. If the connection itself is the limit, the cable change will not show up.
How long does an upgrade last?
That depends on the requirement, not the category. Cat6 specified today covers 1G, 2.5G and 5G at full distance and 10G on short runs, which is a reasonable horizon for most buildings.
Related reading on this site
