Two new desks need wired connections. A small switch, the box that gives you more network sockets, goes under the desk. Then calls freeze, the shared drive disappears and people elsewhere in the office lose their connection too.
One possible cause is a network loop: the new cables have created a circle that lets traffic circulate instead of taking a controlled path. It can disrupt more than the desks you just added. But a protective port shutdown, a bad connection, a power problem or the wrong network settings can produce similar symptoms.
Start with what changed, not a replacement order. Tell your network administrator exactly what was connected and when the trouble began. Adding a switch should not require guessing which cable is safe to pull.
Why an extra connection can take more desks offline
A switch joins wired devices to the network. One connection normally leads back toward the rest of the office. If someone adds a second path without the right design, some traffic can keep circling and multiplying. A broadcast storm is the resulting flood of repeated traffic that can crowd out normal work.
For example, a desk switch might be connected to two wall sockets that both lead back to the same office network. Or an extra cable might join two switches that already have a connection between them. These are possible loop paths, not proof that every pair of cables is wrong.
Planned networks can use redundant connections safely. The difference is that the equipment and settings must control those paths. The manufacturer's guide to network loops and protective port behaviour describes the outage risk and the checks available on supported equipment.
Record the change before trying to fix it
You do not need to diagnose the whole network to give IT a useful starting point. Record:
- When the new switch or cable was connected and when people first noticed trouble.
- Which desks, phones or other work tasks stopped working, including whether WiFi was affected.
- The labels on the wall sockets and the new equipment, without moving any connections.
- Whether the problem affects only internet services or also local resources such as an office file server.
- Any exact warning shown to an authorized administrator, along with its time.
Keep photos and support notes private. Do not include passwords, customer screens or remote access links. A short description such as “the new desk switch was connected to sockets 14 and 15 before the phones dropped” is more useful than “the internet is broken.”
Do not disable loop protection to make a warning disappear. A blocked connection may be keeping the rest of the office online. Let the administrator identify the cause and choose a safe recovery step.
A blocked port is not always a broken port
A port is the socket on the network equipment. STP, short for Spanning Tree Protocol, controls redundant wired paths by keeping a path from forwarding normal traffic when it would create a loop. RSTP is the faster version of that protocol. In a planned redundant layout, seeing one path blocked can be expected.
Other features act differently. Loop protection can disable a port after detecting a loop. A guard setting on a socket intended for a computer can shut that socket down when another switch is connected. Ask which feature acted, not just whether the port is “blocked.”
That distinction affects recovery. The documented loop protection and guard features in the system linked above require manual administrator intervention to reopen a disabled port. Spanning tree instead manages which paths forward traffic. Other equipment may behave differently, so removing a cable does not guarantee that every port will recover on its own.
The administrator should match protection settings to each port's purpose. A connection between switches is not the same as a socket reserved for a staff computer. The switch settings reference explains the supported controls; it is not a universal settings checklist for every brand or model.
Two cables do not automatically mean twice the speed
A spare cable between the same two switches is not an automatic upgrade. With spanning tree, one path may be blocked rather than both carrying ordinary traffic. Without suitable protection, the extra path may create the very outage you wanted to avoid.
Link aggregation is a different option: compatible equipment groups selected physical links into one logical connection. It needs matching support and configuration at both ends. Do not plug in a second cable and assume that grouping will happen, or that one file transfer will become twice as fast. Check the port aggregation requirements and model limitations before choosing that design.
An unmanaged switch is not automatically bad equipment. It may be suitable for a simple approved layout. The question is whether the actual model provides the connection capacity, access controls and loop handling the office needs. Buying a managed switch also does not fix an incorrect design by itself.
The useful fix is knowing why the connection failed, not finding a way to hide the warning.Check more than the new box
The administrator can compare the connection map with event times, inspect affected ports and isolate the newly added part of the network in a controlled way. That may mean removing an unintended path, correcting a port's intended role, replacing a faulty lead or resolving a power issue.
Do not unplug a row of unknown cables or restart everything together. A shared cable may serve phones, cameras or door equipment as well as computers. Agree on the affected systems and a way back before changing a connection. Where work must stop for testing, choose the interruption with the business rather than discovering its scope mid-test.
Wireless links can also matter. An access point, the device providing WiFi in a room, may connect back to the network by cable or through a wireless mesh link. Some combinations can create an unintended second path. The wireless mesh planning guidance warns against enabling a mesh connection on an already wired access point in that system. Do not turn off an office's mesh links blindly; some rooms may depend on them.
If the evidence does not point to a loop, keep investigating. Check the new switch's power and cable connection, and whether its network settings match the approved office setup. An internet-provider fault can coincide with a desk change. The event sequence is a clue, not a verdict.
What a commercial network repair should leave behind
Once the cause is corrected, repeat the work that failed. Check a wired desk, an affected WiFi device, the shared files or business application, and a test call where relevant. Have the administrator review whether fresh port warnings or repeated connection changes appear during the agreed observation period.
Ask for a record of the fault found, the connections changed, final port labels and any settings adjusted. If a protection feature shut a port down, the recovery instructions should say who can restore it and what must be checked first. Do not deliberately recreate a loop on the live office network to prove the fix.
For an expansion or installation, include an agreed go-live support period so issues found during normal work have an owner. A connection diagram should show where the next desks can be added, not just what was plugged in today. Our commercial network handover checklist covers the records worth requesting.
NeuroDesk's commercial network planning and installation starts with the cable routes, equipment and work the office needs to support. For businesses in Windsor, Essex County, London, Chatham and all of Southwestern Ontario, bring the socket labels, the change history and the tasks that failed. We can use those details to plan the investigation before recommending replacement equipment.