In the Data Link Layer of Network Theory, flow control is necessary because a fast sender can inject frames more quickly than a slow receiver can process, buffer, and deliver them upward.2 The Stop-and-Wait protocol is the simplest practical response to this mismatch: the sender transmits one frame, stops, and waits for an acknowledgment before sending the next.2 In a noiseless setting, this provides basic pacing; in a noisy setting, it evolves into Stop-and-Wait ARQ, which adds reliability through acknowledgments, timers, and retransmissions.2

Within the data link layer, protocols may be connectionless or connection-oriented. In connectionless service, each frame is treated independently and often without ordering guarantees beyond the local link behavior; this style is common in many LANs. In connection-oriented service, a logical relationship exists among frames, they are typically numbered, and ordered transfer with acknowledgments becomes meaningful.2 Stop-and-Wait fits naturally into the connection-oriented model because acknowledgments and sequence numbers create state on both sender and receiver.2

Conceptually, Stop-and-Wait can also be viewed as a sliding window protocol with sender window size $W=1$ and receiver window size $W=1$.2 This extreme simplicity makes it easy to implement and reason about, but its chief limitation is poor channel utilization when the propagation delay is large compared with the transmission delay.2

Footnotes

1. Data-link Control & Protocols - Overview of data link layer flow control and stop-and-wait basics. ↩ ↩²

2. Data Link Control Protocols - sandilands.info - Lecture notes giving stop-and-wait throughput and efficiency analysis. ↩ ↩²

3. Simplex Stop and Wait Protocol - University lecture notes relating stop-and-wait to sliding windows and utilization on LAN and satellite links. ↩ ↩² ↩³

4. The Data Link Layer: ARQ Protocols - MIT - Detailed explanation of stop-and-wait ARQ, timeouts, and sequence numbers. ↩

5. III. Link Layer ARQ Protocols - CMU - Notes on stop-and-wait ARQ behavior under losses, delays, and retransmissions. ↩

6. UNIT 2 INTRODUCTION TO DATALINK LAYER - Educational notes distinguishing connectionless and connection-oriented data link protocols. ↩ ↩² ↩³

7. Data Link Layer Protocols - Lecture material linking stop-and-wait to sliding-window operation with window size $1$. ↩ ↩² ↩³

Connectionless vs. Connection-Oriented Data Link Behavior

At the data link layer, a connectionless protocol sends frames as independent units. There is no logical setup phase, no per-exchange state tying one frame to the next, and usually no requirement that frames be numbered for in-order delivery. This design is efficient for many LAN scenarios where the medium is short, error rates are low, and higher layers can handle recovery if necessary.

A connection-oriented protocol, by contrast, establishes logical state before useful transfer, sends frames as part of an ordered exchange, and often ends with a teardown or completion stage. Since the sender and receiver maintain state, mechanisms such as sequence numbers, acknowledgments, and retransmission timers become meaningful and correct.2 Stop-and-Wait and Stop-and-Wait ARQ belong to this stateful family because the sender must know whether the current outstanding frame has been accepted.2

Footnotes

1. UNIT 2 INTRODUCTION TO DATALINK LAYER - Educational notes distinguishing connectionless and connection-oriented data link protocols. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³

2. The Data Link Layer: ARQ Protocols - MIT - Detailed explanation of stop-and-wait ARQ, timeouts, and sequence numbers. ↩ ↩² ↩³

3. Data Link Layer Protocols - Lecture material linking stop-and-wait to sliding-window operation with window size $1$. ↩ ↩²

4. Data-link Control & Protocols - Overview of data link layer flow control and stop-and-wait basics. ↩

Working Mechanism, Sequence Numbers, and Correctness

In the basic noiseless form, the acknowledgment merely signals “ready for the next frame.” However, once errors, loss, or delayed acknowledgments are possible, simple acknowledgments are insufficient.2 The central problem is ambiguity: if a sender times out and retransmits, the receiver must know whether the arriving frame is a new frame or a duplicate of an older one.2

This is why Stop-and-Wait ARQ uses a 1-bit sequence space, commonly $0$ and $1$.2 Since at most one frame can be outstanding, the receiver only needs to distinguish between “the frame I currently expect” and “a duplicate of the one already accepted.” If the sender transmits frame with sequence number $0$, then after correct delivery the receiver sends an acknowledgment indicating it next expects $1$; after accepting frame $1$, it next expects $0$ again.2

This alternating-bit behavior is enough because the sender cannot legally send frame $1$ until frame $0$ is acknowledged, and vice versa.2 Thus, one bit of numbering prevents duplicate delivery while keeping implementation minimal.

Footnotes

1. The Data Link Layer: ARQ Protocols - MIT - Detailed explanation of stop-and-wait ARQ, timeouts, and sequence numbers. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

2. III. Link Layer ARQ Protocols - CMU - Notes on stop-and-wait ARQ behavior under losses, delays, and retransmissions. ↩ ↩²

3. Data Link Layer Protocols - Lecture material linking stop-and-wait to sliding-window operation with window size $1$. ↩ ↩²

4. Simplex Stop and Wait Protocol - University lecture notes relating stop-and-wait to sliding windows and utilization on LAN and satellite links. ↩

Efficiency Analysis and the High-Latency Bottleneck

The principal weakness of Stop-and-Wait is not correctness but efficiency.2 During each cycle, only one frame transmission interval carries useful data, while the sender then idles during forward propagation, acknowledgment return, and reverse propagation. Under the standard simplifying assumptions that ACK transmission time and processing time are negligible, total cycle time is:

$$T_{\text{cycle}} = T_t + 2T_p$$

where $T_t$ is the data-frame transmission time and $T_p$ is the one-way propagation delay.2

Hence link utilization or efficiency is approximately:

$$\eta = \frac{T_t}{T_t + 2T_p}$$

If we define

$$a = \frac{T_p}{T_t},$$

then

$$\eta = \frac{1}{1 + 2a}$$

2

This formula explains the bottleneck clearly. When $a \ll 1$, the frame transmission time dominates and utilization is respectable. When $a \gg 1$, propagation dominates, so the sender spends most of its time waiting rather than sending.2 This is especially harmful on long-distance, satellite, or very high-bandwidth links, where frames are launched quickly but acknowledgments return slowly.2

For example, if $a=10$, then:

$$\eta = \frac{1}{21} \approx 0.0476$$

so less than $5\%$ of the time carries useful payload. Some lecture material notes that for satellite links, utilization may drop to around $0.001$ or lower under certain parameters, illustrating severe bandwidth waste.

Footnotes

1. Data Link Control Protocols - sandilands.info - Lecture notes giving stop-and-wait throughput and efficiency analysis. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

2. Simplex Stop and Wait Protocol - University lecture notes relating stop-and-wait to sliding windows and utilization on LAN and satellite links. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

Lost Frames, Lost ACKs, and Delayed ACKs

Stop-and-Wait ARQ is designed specifically for noisy or unreliable links where either data frames or acknowledgments may be lost, damaged, or delayed.2

1. Lost data frame

If the sender transmits a frame and that frame never reaches the receiver, no acknowledgment can be generated. Without a timer, the sender would wait forever. Therefore, timeout is essential: once the timer expires, the sender retransmits the same frame.2

2. Lost acknowledgment

If the receiver gets the frame correctly and sends the ACK, but the ACK is lost, the sender again sees silence and eventually retransmits.2 Now the receiver encounters a duplicate frame. Because the duplicate has the old sequence number, the receiver discards the payload and sends the ACK again rather than delivering the same packet twice.2

3. Delayed acknowledgment

A delayed ACK is more subtle. The sender may time out and retransmit before the original ACK arrives. When that late ACK eventually appears, the sender must not confuse it with permission for a future frame. Sequence-numbered acknowledgments prevent this ambiguity by tying the ACK to the expected next frame number.2

This recovery method is graceful because no complex rollback is required. The protocol uses only three ideas: detect absence of confirmation, retransmit after timeout, and reject duplicates using sequence numbers.2

Footnotes

1. The Data Link Layer: ARQ Protocols - MIT - Detailed explanation of stop-and-wait ARQ, timeouts, and sequence numbers. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

2. III. Link Layer ARQ Protocols - CMU - Notes on stop-and-wait ARQ behavior under losses, delays, and retransmissions. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

3. Data Link Layer Protocols - Lecture material linking stop-and-wait to sliding-window operation with window size $1$. ↩

Strengths, Limitations, and Place in the Data Link Layer

Stop-and-Wait remains pedagogically and practically important because it isolates the core ideas behind reliable data transfer: pacing, acknowledgment, timeout, retransmission, and duplicate detection.2 Its implementation cost is low, buffer needs are minimal, and correctness is comparatively easy to verify.2

However, its limits are fundamental:

• only one outstanding frame is allowed, so the link is often idle2
• throughput collapses as round-trip propagation delay grows2
• performance becomes poor on high bandwidth-delay product links2
• even when errors are rare, waiting for every ACK wastes capacity2

Thus, Stop-and-Wait is best understood as the conceptual foundation for more capable ARQ schemes such as Go-Back-N and Selective Repeat, which keep multiple frames in flight to improve utilization.2

A compact way to summarize the protocol is:

$$\text{Correctness} \approx \text{ACKs} + \text{timeout} + \text{sequence numbers}$$

while its main performance issue is:

$$\text{Low utilization when } 2T_p \gg T_t$$

2

Footnotes

1. Data-link Control & Protocols - Overview of data link layer flow control and stop-and-wait basics. ↩

2. The Data Link Layer: ARQ Protocols - MIT - Detailed explanation of stop-and-wait ARQ, timeouts, and sequence numbers. ↩ ↩² ↩³

3. Data Link Layer Protocols - Lecture material linking stop-and-wait to sliding-window operation with window size $1$. ↩

4. Data Link Control Protocols - sandilands.info - Lecture notes giving stop-and-wait throughput and efficiency analysis. ↩ ↩² ↩³ ↩⁴ ↩⁵

5. Simplex Stop and Wait Protocol - University lecture notes relating stop-and-wait to sliding windows and utilization on LAN and satellite links. ↩ ↩² ↩³ ↩⁴ ↩⁵

6. III. Link Layer ARQ Protocols - CMU - Notes on stop-and-wait ARQ behavior under losses, delays, and retransmissions. ↩