How Data Link Layers Provide Flow and Error Control (with Examples)

The Data Link Layer provides two key services to make transmission dependable and efficient on a shared physical medium: flow control and error control.

At a high level:

• Flow control coordinates sender/receiver rates so the receiver’s buffering is not exceeded.
• Error control ensures corrupted frames are detected (usually via a checksum) and retransmitted when needed (via ARQ-style protocols).

Below is a structured view of the dominant approaches used in link layers, including concrete examples (e.g., stop-and-wait, sliding window, CRC-based detection, and retransmission strategies).

Note: The exact flow/error control used depends on the link technology (e.g., Ethernet vs. Wi-Fi) and link-layer protocol (e.g., HDLC/LLC-like framing, or 802.11 MAC reliability mechanisms).

Key building blocks: frames, checksums, and acknowledgements

Most data link protocols follow a common pattern:

1. Frame formation
   The sender groups bits into a frame with fields like header, payload, and a FCS (or checksum).

2. Error detection
   The receiver recomputes the CRC/checksum over the received frame and compares it to the FCS / checksum.

3. Error recovery (ARQ) If the check fails, the receiver triggers ARQ behavior—commonly by sending ACK (success) or NACK/reject (failure) and the sender retransmits.

This is conceptually similar to transport-layer reliability, but it is done hop-by-hop (per link), so the link layer can isolate local corruption and delay retransmission to the right place.

Key terms to remember

• Frame
• Flow control
• Error control
• ARQ
• FCS

From stop-and-wait to sliding window (improves throughput)

To increase efficiency, many link layers use a sliding window protocol with sequence numbers.

• Sender may transmit multiple frames before receiving all ACKs.
• Receiver acknowledges correctly received frames (cumulatively and/or selectively).
• On error detection, the sender retransmits based on the acknowledgment policy.

Common ARQ strategies

• Go-Back-N
• Selective Repeat

These strategies determine how much retransmission overhead occurs when corruption affects a specific frame.

Flow control in the data link layer: preventing buffer overflow

Even if frames are error-free, the receiver can become overloaded due to:

• variable processing time,
• limited buffer space,
• slower consumption by upper layers,
• bursts from a fast sender.

Data link flow control mitigates this with one (or a combination) of:

1. Window-based flow control
   The sender’s maximum number of outstanding frames is limited by a receiver-advertised or negotiated window size. This naturally caps the data rate to what the receiver can store.

2. Receiver-driven backpressure
   The receiver signals “slow down” (or “stop”) using MAC/control frames or by reducing allowed credit.

3. Pause/feedback signals (technology-specific)
   Some technologies include explicit mechanisms to temporarily reduce transmission rate.

Putting it together: example scenarios

Example 1: Noisy link with stop-and-wait ARQ

Assume frames get corrupted with some probability $p$. With stop-and-wait:

• corruption causes receiver to reject or not ACK,
• sender times out and retransmits.

Flow is inherently safe: only one frame is outstanding.

Where error control acts: at the receiver’s CRC/FCS check, then via retransmission triggered by ACK/NACK or timeout.

Where flow control acts: by allowing only one in-flight frame.

Example 2: High-latency link using sliding window + selective repeat

With longer propagation delay, stop-and-wait wastes time waiting for ACKs. Sliding window allows the sender to keep the link busy.

Selective repeat:

• receiver accepts good frames even if some earlier ones are missing (within the window),
• sender retransmits only the missing/corrupted frames.

Flow control mechanism: window size bounds outstanding frames to receiver buffer capacity.

Error control mechanism: per-frame CRC/FCS detection + targeted retransmissions.