latency
Latency is the time lag between asking for something and the response beginning — how long a bit takes to get from here to there. If bandwidth is how much you can send and throughput is how much actually arrives per second, latency is purely about time: the delay. It is what makes a video call feel laggy or a game feel snappy, and it is felt even when bandwidth is plentiful.
People measure it two ways. One-way latency is the time for a bit to travel from sender to receiver. More commonly quoted is round-trip time (RTT) — the time for a small message to go to the other end and a reply to come back, which is roughly what ping reports. Total latency on a path is the sum of the four delay components at each hop: transmission delay (L/R, clocking the packet onto the link), propagation delay (distance over signal speed), queuing delay (waiting in router buffers), and processing delay (a router inspecting the header). On long paths, propagation usually dominates; under congestion, queuing does.
The headline truth: latency and bandwidth are independent, and you cannot buy your way past the speed of light. Doubling your bandwidth does nothing for the round trip to a distant server — that floor is set by distance. This is why a 'fast' (high-bandwidth) connection can still feel slow for interactive work over long distances, and why low latency, not just high bandwidth, is the goal for gaming, video calls, trading, and any back-and-forth conversation.
Ping a server and it reports an RTT of, say, 80 ms — that is the round-trip latency. Upgrading from 100 Mbps to 1 Gbps will not change that 80 ms one bit; only a shorter path or less queuing will.
Delay, not capacity — measured one-way or as round-trip time, and floored by distance.
Latency, bandwidth, and throughput are three different things, and confusing them is the single most common networking mistake. More bandwidth never reduces latency; only shortening the path or relieving congestion does.