Flow Control and Error Control

Imagine two devices sending data to each other over a single connection — say your laptop talking to the Wi-Fi router. Two things can go wrong. First, the sender might push data out faster than the receiver can handle, and some of it gets dropped. Second, the data itself might get scrambled or lost somewhere along the way. Flow control is the set of rules that fixes the first problem (pacing), and error control is the set of rules that fixes the second (correctness). This lesson is the map for the whole topic — we'll meet both ideas at a high level, then dig into each protocol in its own lesson.

Throughout this topic, the chunk of data being sent is called a frame. You don't need to know exactly how a frame is built yet — just picture it as one labelled envelope of data that travels from the sender to the receiver.

Every receiver has a small holding area in memory called a buffer where incoming data waits to be processed. That buffer is not infinite. If the sender fires data faster than the receiver can empty its buffer, the buffer fills up and any further data is simply dropped — wasted effort that has to be sent all over again. Flow control gives the receiver a way to say 'slow down' or 'okay, send more,' so the two sides stay in step.

Even when the pace is right, data can still go wrong in three ways: a frame can arrive corrupted (some bits flipped by electrical noise), it can be lost entirely (never arrives), or it can be duplicated (the same frame arrives twice). Error control catches these problems and repairs them. The usual fix is simple: the receiver confirms what it got, and anything missing or broken gets sent again.

The whole 'confirm and re-send' idea has a name: ARQ. The receiver sends back a short message called an acknowledgment (ACK) to say 'got it.' The sender starts a timer each time it sends; if the acknowledgment doesn't come back before the timer runs out, the sender assumes the frame was lost and re-sends it. To tell frames apart — and to spot duplicates — each frame carries a sequence number.

These four tools show up again and again across every protocol in this topic, so it's worth getting comfortable with them now:

Flow control and error control aren't done in just one place. They run at the data link layer — between two devices connected by a single hop, such as your laptop and the router — where the data unit is the frame. The very same ideas run again at the transport layer, where a protocol called TCP (Transmission Control Protocol) provides flow control and error control from end to end, all the way across the internet. Same concepts, different scale.

The rest of this topic walks through three protocols that combine flow control and error control. They differ mainly in one thing: how many frames the sender is allowed to send before it has to stop and wait for a confirmation. Here they are at a glance — we'll unpack each one in its own lesson: