Wi-Fi collapse in a Maltese secondary school computer lab
A Church secondary school in Birkirkara has 320 students and 45 desktop PCs. Its 30 Mbps fibre link is shared by administration and two computer labs. A new class set of 24 Wi-Fi laptops was introduced this term. Teachers report that when more than about 18 laptops connect in Room 2, logins to the local domain controller take over four minutes and several laptops drop the wireless connection entirely. The access point is a single-band 802.11n model mounted in the corridor outside Room 2; it is six years old and broadcasts on channel 6, the same channel used by two neighbouring schools' access points. Fixed desktops in the same room use Ethernet and log in normally. The head of school has €4,000 and has been advised by a supplier to 'upgrade the internet to 100 Mbps'.
1. Using the evidence in the scenario, identify the most likely bottleneck and justify your choice. State what additional measurement you would take before spending any money.
Read the worked answer
The evidence points to the wireless access point and the 2.4 GHz band, not the internet link.
- Fixed Ethernet PCs log in normally. Domain logins are handled by the local domain controller, so they do not traverse the 30 Mbps fibre link. If the fibre were the bottleneck, wired and wireless clients would both be slow.
- The failure threshold is around 18 clients. That is a contention symptom: a single 802.11n access point on one 2.4 GHz channel shares airtime among all clients, and management/authentication frames add overhead. Adding clients increases collisions and retries, which shows up as dropped connections and long waits.
- Channel 6 is shared with two neighbouring schools. Co-channel interference from outside the building cannot be fixed by the school's own equipment settings alone, but it explains why performance is worse at certain times.
- The AP is in a corridor. Wall attenuation and distance reduce signal quality, so clients fall back to lower data rates, consuming more airtime per login.
Before spending, I would run a site survey: measure signal strength (RSSI) and signal-to-noise ratio in Room 2, list visible SSIDs and channels, and test logins with a wired laptop on the same switch. I would also check the domain controller's CPU and disk during login storms. This distinguishes an RF problem from a server problem.
2. Explain why moving to a 100 Mbps internet connection would not, on its own, solve the login delays and dropped connections.
Read the worked answer
Internet bandwidth is only used for traffic that leaves the school network. Authentication against the local domain controller, loading roaming profiles, and mapping home folders all happen on the LAN. The 30 Mbps figure describes the WAN uplink; the bottleneck described is airtime contention on a single 2.4 GHz access point.
A useful analogy is a wide motorway exit feeding a single narrow toll booth: widening the motorway does not increase the number of cars passing the booth per minute. Similarly, increasing the fibre speed changes nothing about how many laptops can share one radio channel.
Two further points make the upgrade poor value here:
- The symptom is drops, not slow downloads. Dropped wireless connections are an RF/association problem, not a throughput-ceiling problem.
- No baseline has been measured. Without a test showing the 30 Mbps link saturated during logins, the supplier's claim is unsupported. The school could pay for 100 Mbps and see no change in Room 2.
The money would be better spent on a dual-band 802.11ac access point (or two), 5 GHz channel planning, and Ethernet cabling to a second AP location.
3. Propose a spending plan within the €4,000 budget. Justify the priorities and note one trade-off the school must accept.
Read the worked answer
Priority 1 - fix the radio problem (about €700–€1,200). Replace the corridor AP with two dual-band 802.11ac access points, one inside Room 2 and one covering the adjacent lab. Use 5 GHz for the laptops, set 20/40/80 MHz channels sensibly, and disable legacy 802.11b rates. Two APs reduce the number of clients per radio and shorten the distance through walls.
Priority 2 - backhaul (about €400–€900). Access points must be wired to the switch with Cat6 and powered by PoE. A wireless repeater would halve throughput and repeat the original mistake.
Priority 3 - verify (about €100–€300). A simple site survey app and a loaned laptop to test before and after. Without a before-and-after measurement the school cannot prove the spend worked.
Priority 4 - keep a contingency (about €500). If the survey shows the domain controller is also struggling, some budget may be needed for RAM or an SSD.
Trade-off: the school accepts that 5 GHz has shorter range than 2.4 GHz, so coverage may need a third AP if the survey shows weak signal at the far end of the room. Also, two APs need two cable runs, which may require an electrician and cannot be done during lesson time. Spending the whole budget on internet bandwidth would leave the actual fault untouched.