Resource Category
Network performance, explained
Network performance is the end-to-end behavior of a connection, not the spec sheet of any single component. The slowest link in the chain — your Wi-Fi, your ISP, an upstream peer, or the destination server — sets the experience.
How data travels across the internet
Every request crosses three roughly distinct zones: your local network (device → Wi-Fi → router), your ISP's access and backbone networks, and the destination's edge and origin infrastructure. Between ISPs, traffic moves through peering points and transit providers. Each hop adds latency and is a potential failure point.
Throughput vs bandwidth
Bandwidth is the theoretical maximum capacity of a link. Throughput is what you actually achieve. Throughput is always lower than bandwidth — protocol overhead, retransmissions, contention, and TCP's congestion control all eat into it. A 1 Gbps connection rarely delivers 1 Gbps of useful application data.
Common bottlenecks
- Wi-Fi — by far the most common bottleneck in homes and small offices.
- The router itself — older or under-powered routers cap real throughput.
- ISP last-mile congestion — neighborhood saturation during peak hours.
- Peering disputes — your ISP and the destination's network may have a slow path between them.
- The origin server — slow backends look like slow internet.
Network congestion
Congestion happens whenever demand exceeds a link's capacity. TCP responds by slowing senders down, so congestion shows up as both lower throughput and higher latency at the same time. Quality of Service (QoS) can prioritize sensitive traffic (calls, gaming) over bulk transfers, partly mitigating the symptoms.
Routing
Internet routing decides which path packets take across the network. BGP — the protocol that does this — optimizes for policy and reachability, not always for speed. Sometimes a packet from New York to Boston travels through Chicago because that's the cheapest path between the involved networks.
How to think about diagnosing slowness
Start local (Ethernet test vs Wi-Fi), then check the immediate upstream (modem, ONT), then trace the path to confirm where latency jumps. A traceroute that shows a 200 ms hop in the middle of the path is pointing at the problem hop. See the network diagnostics guide for the full workflow.