Sliding Window Protocols — Architecture & Dynamics

Overview

Sliding window protocols regulate how many packets a sender may transmit before requiring acknowledgment. They provide a structured way to manage ordering, reliability, and throughput in communication systems. By maintaining a window of outstanding, unacknowledged packets, these protocols balance link utilization with receiver capacity and network conditions.

Sliding window mechanisms are foundational in transport protocols (TCP), link‑layer ARQ systems, and wireless MACs, and they complement both ARQ and flow control mechanisms.

Architectural Principles

Window Concept

The sender maintains a window of sequence numbers representing packets that may be transmitted but are not yet acknowledged. The window “slides” forward as ACKs arrive.

A window is defined by:

• Lower edge — the oldest unacknowledged packet
• Upper edge — the highest sequence number allowed for transmission
• Window size — the number of packets allowed in flight

Sender Responsibilities

The sender must:

• transmit packets within the window
• buffer unacknowledged packets
• retransmit lost or corrupted packets
• advance the window when ACKs arrive

Receiver Responsibilities

The receiver must:

• detect missing or out‑of‑order packets
• buffer packets (Selective Repeat) or discard them (Go‑Back‑N)
• generate ACKs or NACKs
• maintain ordering for delivery to upper layers

Window Types

Fixed Window

The window size is constant and determined by protocol parameters or buffer constraints.

Dynamic Window

The window size adapts to:

• network congestion
• receiver capacity
• round‑trip time
• error rate

TCP uses a dynamic window through congestion control and flow control mechanisms.

Core Sliding Window Variants

Stop‑and‑Wait

A degenerate sliding window with size 1.

• sender transmits one packet
• waits for ACK
• retransmits on timeout

Pros: simple, minimal state
Cons: extremely low throughput on high‑latency links

Go‑Back‑N (GBN)

A sender may transmit up to N unacknowledged packets.

• receiver accepts packets in order only
• on error, sender retransmits from the missing packet onward
• cumulative ACKs

Pros: simple receiver
Cons: inefficient when error rate is high

Selective Repeat (SR)

The most efficient sliding window variant.

• receiver accepts packets out of order
• only missing packets are retransmitted
• requires per‑packet ACKs or selective ACKs
• requires reordering buffers

Pros: high throughput, minimal retransmission
Cons: more complex state machines and larger buffers

Window Dynamics

Window Advancement

The window slides forward when:

• ACKs arrive
• timeouts expire (triggering retransmission)
• receiver advertises new buffer space

Sequence Number Space

Sequence numbers must be large enough to avoid ambiguity between:

• new packets
• retransmissions
• wrapped sequence numbers

Selective Repeat requires a sequence space at least twice the window size.

ACK Strategies

• Cumulative ACKs — acknowledge all packets up to a sequence number
• Selective ACKs (SACK) — explicitly acknowledge non‑contiguous packets
• NACKs — request retransmission of specific packets

Performance Considerations

Throughput

Throughput depends on:

Larger windows improve throughput, especially on high‑latency links.

Latency

Latency is influenced by:

• retransmission delays
• ACK frequency
• window size
• buffering at sender and receiver

Efficiency

Selective Repeat is the most efficient under moderate to high error rates.
Go‑Back‑N is simpler but wastes bandwidth on retransmissions.

Integration with ARQ and Flow Control

With ARQ

Sliding window protocols define when to retransmit.
ARQ defines what to retransmit and how errors are detected.

With Flow Control

Flow control (credit‑based or rate‑based) limits how much data can be in flight.
Sliding windows limit how many unacknowledged packets can be in flight.

Together they ensure:

• no buffer overflow
• no congestion collapse
• ordered, reliable delivery

Use Cases in Real Systems

TCP

TCP uses a dynamic sliding window with:

• cumulative ACKs
• selective acknowledgments (SACK)
• congestion control (slow start, congestion avoidance)
• flow control (receiver‑advertised window)

TCP’s sliding window is central to Internet reliability and performance.

Wireless MACs

Wi‑Fi and cellular systems use sliding windows to manage:

• variable channel quality
• retransmissions
• out‑of‑order reception

Often combined with HARQ for improved robustness.

Link‑Layer Protocols

Many link‑layer ARQ systems (e.g., HDLC, PPP extensions) use Go‑Back‑N or Selective Repeat.

Comparison of Sliding Window Variants

Variant	Out‑of‑Order Reception	Retransmission Strategy	Complexity	Efficiency
Stop‑and‑Wait	No	Retransmit single packet	Very Low	Very Low
Go‑Back‑N	No	Retransmit from error onward	Low	Medium
Selective Repeat	Yes	Retransmit only missing packets	Medium‑High	High

Design Tradeoffs

• Window Size vs Memory — larger windows require more buffering.
• Sequence Space vs Complexity — Selective Repeat requires larger sequence spaces.
• Latency vs Throughput — larger windows improve throughput but may increase latency.
• Error Rate vs Efficiency — Selective Repeat excels in noisy channels.

Related Pages

• ARQ — Automatic Repeat Request
• Credit‑Based Flow Control — Architecture & Use Cases
• Protocol Flow Control — Architecture & Mechanisms
• Retransmission Strategies — Architecture & Tradeoffs
• Link Layer Reliability Mechanisms

Summary

Sliding window protocols regulate how many packets may be in flight, balancing throughput, reliability, and ordering. They are fundamental to modern communication systems, from TCP to wireless MACs, and they integrate tightly with ARQ and flow control mechanisms to ensure efficient, reliable data transfer.