The Intermittent Network Outage
A company has a network with 50 computers connected to a 48-port switch. The switch is connected to a router that provides internet access. Users report that every day around 2 PM, the network becomes very slow for about 10 minutes, and sometimes connections drop. The IT support technician checks the switch and notices that the link light for port 24 is blinking rapidly during these times. Port 24 connects to a server that runs a backup job every day at 2 PM, backing up data to a cloud service. The server has a 1 Gbps network card, and the switch ports are 1 Gbps. The technician also observes that the switch's CPU utilization spikes to 90% during the slowdown. The network uses CAT5e cables. The backup job transfers about 500 GB of data each day. The company has a 100 Mbps internet connection. The technician suspects a network loop but has not found one. The switch is unmanaged.
1. What is the most likely cause of the network slowdown, and what evidence supports this?
Read the worked answer
The most likely cause is network congestion due to the backup job saturating the internet link. The backup transfers 500 GB to the cloud over a 100 Mbps connection. At 100 Mbps, 500 GB takes about 11 hours (500 GB = 4000 Gb; 4000 Gb / 100 Mbps = 40000 seconds ≈ 11.1 hours). However, the slowdown lasts only 10 minutes, so the backup might be using compression or the transfer might be bursty. But the key is that the backup runs at 2 PM, coinciding with the slowdown. The switch CPU spike to 90% suggests high traffic volume. The link light on port 24 blinking rapidly indicates heavy traffic from the server. The unmanaged switch has no VLANs or QoS, so the backup traffic competes with regular traffic. The 100 Mbps internet is the bottleneck; the backup saturates it, causing slow internet for others. Also, the switch might be overloaded if the backup server is sending data at 1 Gbps to the switch, but the switch's uplink to the router might be 1 Gbps, so the bottleneck is the internet. The fact that connections drop could be due to buffer overflow on the router or switch. The technician should check the router's WAN utilization during the slowdown. Also, the backup job might be using multiple connections, exacerbating congestion. The evidence: timing correlation, high switch CPU, port 24 activity, and known backup schedule.
2. What steps should the technician take to confirm the cause and mitigate the issue without upgrading the internet connection?
Read the worked answer
To confirm: monitor the router's WAN interface utilization during the slowdown using SNMP or the router's logs. Check the switch's port statistics for port 24 (if managed, but it's unmanaged, so maybe use a network tap or mirror port). Alternatively, run a bandwidth monitor on the server or use Windows Performance Monitor to see network throughput. Also, check if the backup job can be throttled. To mitigate without upgrading internet: (1) Schedule the backup for off-hours (e.g., 2 AM) when network usage is low. (2) Implement QoS on the router to prioritize business traffic over backup traffic, if the router supports it. (3) Throttle the backup software's bandwidth usage. (4) Use deduplication or compression to reduce data transfer. (5) Consider a local backup first, then cloud sync during off-hours. (6) If the switch is unmanaged, consider replacing with a managed switch to enable VLANs and QoS, but that's a cost. The simplest is rescheduling. Also, check if the backup is incremental; if full backup daily, change to incremental. The technician should also verify that the 100 Mbps connection is actually the bottleneck by testing internet speed during non-backup times.