Network Delays in Computer Networks + End-to-End Delay Calculation (200 MB over 1 Gbps with 3 Routers)

In computer networks, the time a message takes to travel from a source host to a destination host is called network delay (often discussed as latency). A packet experiences multiple components at each hop (router-to-router or host-to-router), commonly modeled as:

keyword Transmission delay

keyword Propagation delay

keyword Queuing delay

keyword Processing delay

A standard breakdown is that nodal delay equals the sum of processing, queuing, transmission, and propagation delays at each router, and end-to-end delay accumulates these across hops. 2

Types of Delays (What they mean and why they matter)

1) Processing delay ($d_{proc}$)

This is the time a router (or switch) takes to process the packet header—e.g., checking bit-level errors, determining the outgoing interface, updating header fields, etc.

2) Queuing delay ($d_{queue}$)

When the router is busy transmitting other packets, an arriving packet must wait in a queue. The queuing delay depends on traffic intensity/congestion and buffer state. 2

3) Transmission delay ($d_{trans}$)

Transmission delay is the time required to put all bits of the packet onto the link. If the packet has length $L$ bits and the link transmits at rate $R$ bits/s (bandwidth in this context), then:

• $d_{trans}=\frac{L}{R}$ This is sometimes called the “store-and-forward delay” component for packet switching analysis. 2

4) Propagation delay ($d_{prop}$)

Propagation delay is the time for the signal to travel from one node to another through the transmission medium. If distance is $d$ and propagation speed is $s$, then:

• $d_{prop}=\frac{d}{s}$ It depends on distance and medium speed, not directly on packet size. 2