Network Fundamental - Summary

1. Internet Structure & Network Fundamentals

1.1 Protocol

A protocol is a set of rules for how data is sent and received over a network.
Examples: HTTP, TCP, IP, FTP.

1.2 Internet Structure

Key Components:

Network Edge: Devices like clients and servers (e.g., phones, laptops, web servers).
Access Networks & Physical Media: Connect users to the internet using DSL, cable, fiber, or wireless.
Network Core: Routers and links that connect different networks using packet switching.

1.3 Access Networks

DSL (Digital Subscriber Line)

Uses telephone lines to send data and voice at the same time.
Speeds: 1–25 Mbps.
DSL modem separates voice and data signals.

HFC (Hybrid Fibre-Coaxial)

Combines fiber-optic and coaxial cables.
Uses encryption and Frequency Division Multiplexing (FDM) to separate users.
Speeds: Up to hundreds of Mbps.

NBN (National Broadband Network)

Australia’s national network.
Expensive and often slow (around 55 Mbps).
FTTH (Fibre to the Home) brings fiber directly to homes and businesses.

Enterprise Networks

Private high-speed networks for companies.
Uses dedicated fiber lines.

1.4 Physical Media

Bit: Smallest unit of data.
Physical Link: Connection between devices.
- Guided Media (Wired): Copper, coaxial, fiber.
- Unguided Media (Wireless): Air, radio waves.

1.5 Packet Switching & Delays

Data is split into packets.
Routers forward packets one hop at a time.

Types of Delay:

Transmission Delay:
Delay = L / R
(L = packet size, R = link speed)
Store and Forward:
Routers must receive the full packet before sending.
End-to-End Delay:
Total Delay = L / R1 + L / R2
Queueing Delay:
Happens when too many packets arrive at once.
Can cause buffering or packet loss.

1.6 Network Core Functions

Function	Description
Routing	Decides the full path across the network
Forwarding	Chooses next router to send the packet

1.7 Circuit Switching

Reserves full path before sending data.
Guarantees performance but wastes resources if idle.

Multiplexing:

FDM (Frequency Division Multiplexing):
Each user gets a different frequency.
TDM (Time Division Multiplexing):
Each user gets a time slot.

Type	Best for	Description
FDM	Analog systems	Continuous transmission
TDM	Digital systems	More efficient and flexible

1.8 Packet Switching vs Circuit Switching

Feature	Circuit Switching	Packet Switching
Bandwidth Usage	Reserved (may be idle)	Shared (efficient)
Scalability	Lower	Higher
Example (2 Mbps)	2 users max	3+ users possible

1.9 Internet Structure

The internet is a "network of networks".
Each organization (ISP, company, university) manages its own part of the internet.

2. Networking Performance & Internet Stack

2.1 Sources of Packet Delay

Transmission Delay

Time to push all bits into the link.
Formula:
$\text{Transmission Delay} = \frac{L}{R}$
Where L = packet length (bits), R = link bandwidth (bits per second)

Propagation Delay

Time for a signal to travel through the medium.
$\text{Propagation Delay} = \frac{\text{distance}}{\text{propagation speed}}$

Queueing Delay

Time a packet waits in the queue until it can be transmitted.
Depends on traffic load.

Processing Delay

Time to examine packet header and determine forwarding.

2.2 Traffic Intensity (La/R)

La/R ≈ 0 → Minimal queueing delay
La/R → 1 → Large queueing delay
La/R > 1 → Queue builds indefinitely; infinite delay

2.3 Packet Loss

Occurs when the queue (buffer) is full.

Incoming packets are dropped
May be retransmitted by sender, router, or lost permanently

2.4 Throughput

Throughput is the rate at which bits are successfully delivered from sender to receiver.

Instantaneous Throughput: At a specific moment
Average Throughput: Over a time interval
$\text{Throughput} = \frac{\text{Total bits received}}{\text{Total time}}$

2.5 Internet Protocol Stack (TCP/IP)

The 5-layer model of the Internet:

Application Layer – Network applications (e.g., HTTP, FTP, DNS)
Transport Layer – Process-to-process data transfer (e.g., TCP, UDP)
Network Layer – Routing of datagrams (e.g., IP)
Link Layer – Data transfer between adjacent network nodes (e.g., Ethernet, WiFi)
Physical Layer – Transmission of raw bits (e.g., copper, fiber, radio)

2.6 Packet Overlap & Queue Test (Visual Method)

Use color-coded diagrams to check for overlap in packet transmission.

If the red packet finishes after the green starts → overlap (queuing occurs)
If the red finishes before the green starts → no overlap

Visual Method

Compute each packet’s transmission time using $\frac{L}{R}$
Compare end time of earlier packet to start time of next
Helps identify if traffic will queue or pass freely

2.7 TCP/IP Model Summary Table

Application Layer: Web, Email, FTP, DNS
Transport Layer: TCP, UDP
Network Layer: IP, ICMP, Routing
Link Layer: Ethernet, Wi-Fi, PPP
Physical Layer: Copper, Fiber, Radio

3. Application Layer (1 of the 5 Layers)

3.1 What an Application-Layer Protocol Defines

An application-layer protocol defines:

Types of messages exchanged (e.g., request, response)
Message syntax: Format and structure
Message semantics: Meaning of info in each field
Rules for sending/receiving messages

Protocol Examples

Open Protocols (e.g., HTTP, SMTP):
- Publicly defined in RFCs
- Interoperable across systems
- Example: Google Chrome uses HTTP
Proprietary Protocols (e.g., Zoom, MS Teams):
- Privately owned and maintained

3.2 Internet Transport Protocol Services

Protocol	Features
TCP	Reliable, ordered delivery, congestion control, connection-oriented
UDP	Unreliable, faster, minimal overhead, no connection setup

Note: TCP is better for accuracy (e.g., email, downloads),
UDP is faster for streaming (e.g., video calls), but may need buffering.

Example Transport Use by Application

Application	Protocol Used
FTP (File Transfer)	TCP
SMTP (Email)	TCP
HTTP (Web)	TCP
Skype, Zoom	UDP or TCP
Streaming Video	UDP or TCP

3.3 Web and HTTP

What is a Web Page?

A web page consists of:

Objects: Images, text, audio, video
Base HTML file and additional objects
Each object is addressable via a URL
Example:
www.somesite.com/images/logo.png

HTTP: Stateless Protocol

The server does not remember past requests
Each HTTP request is independent

Persistent vs Non-Persistent HTTP

Type	Description
Persistent	Multiple objects sent over single TCP connection (faster, preferred)
Non-Persistent	One object per TCP connection (more overhead)

Tip: Use persistent HTTP for:

Web browsing
API requests
Efficient resource fetching

4. Network Fundamentals

4.1 TCP and UDP

TCP (Transmission Control Protocol): Reliable, ensures data arrives in order and without errors.
UDP (User Datagram Protocol): Unreliable but faster, no error checking or ordering.

Example: Python UDP Server

from socket import *
serverPort = 12000
serverSocket = socket(AF_INET, SOCK_DGRAM)
serverSocket.bind(("", serverPort))
print("The server is ready to receive")

while True:
    message, clientAddress = serverSocket.recvfrom(2048)
    modifiedMessage = message.decode().upper()
    serverSocket.sendto(modifiedMessage.encode(), clientAddress)

4.2 Cache and Cookies

Cache: Temporarily stores website content for faster loading on revisits.
Cookies: Store user-specific data (login status, preferences, shopping carts).

Analogy:

Cache = remembers what you've seen (e.g., layout).
Cookies = remember who you are (e.g., login).

4.3 DNS (Domain Name System)

How DNS Works:

Computer contacts Local DNS Server.
If unknown, Local DNS contacts Root Server.
Root → TLD Server → Authoritative Server.
IP is returned to your computer to load the site.

4.4 Email Protocols

SMTP (Simple Mail Transfer Protocol)

Used for sending emails.
Works between clients and mail servers or between servers.

POP3 vs. IMAP

POP3: Downloads and removes mail from the server.
IMAP: Keeps mail on the server; allows syncing across devices.

Email Journey (Alice to Bob)

Alice writes and sends via SMTP.
Travels server to server using SMTP.
Bob retrieves using IMAP.
Bob replies; Alice retrieves via POP3.

SMTP Response Codes

2xx = Success (e.g., 250 OK)
3xx = Intermediate (e.g., 354 Start mail input)
4xx = Temporary error (e.g., 450 Mailbox unavailable)
5xx = Permanent error (e.g., 550 User does not exist)

SMTP vs HTTP Message Endings

SMTP: Ends message with a single . line.
HTTP: Uses Content-Length header.

4.5 Python TCP Server Example

from socket import *
serverPort = 12000
serverSocket = socket(AF_INET, SOCK_STREAM)
serverSocket.bind(('', serverPort))
serverSocket.listen(1)
print("The server is ready to receive")

while True:
    connectionSocket, addr = serverSocket.accept()
    sentence = connectionSocket.recv(1024).decode()
    capitalizedSentence = sentence.upper()
    connectionSocket.send(capitalizedSentence.encode())
    connectionSocket.close()

Explanation

socket(AF_INET, SOCK_STREAM): IPv4, TCP.
bind(): Assign IP and port.
listen(): Queue connections.
accept(): Wait for client.
recv() / send(): Receive and send data.
close(): Close the connection.

4.6 Encoding vs Decoding

Bytes: Machine-readable data (used in transport).
Strings: Human-readable text.
ASCII vs Unicode (UTF-8): ASCII for 127 characters; Unicode supports over 1 million.

Use encode() to convert string to bytes. Use decode() to convert bytes to string.

5. Transport Layer

5.1 UDP – User Datagram Protocol

UDP sends data without ensuring delivery.
No handshake or error checking.
Used for fast, real-time apps (e.g., games, VoIP).

5.2 TCP vs UDP

TCP is reliable and connection-based (used by HTTP).
UDP is faster but unreliable.

5.3 Port Numbers

Identify apps on a device.
Ranges:
- 0–1023: Well-known (HTTP 80, FTP 21)
- 1024–49151: Registered
- 49152–65535: Dynamic/Temporary

5.4 Checksum (UDP)

Detects data errors using 1’s complement.
Sender adds data, flips bits → sends checksum.
Receiver re-adds and checks for all 1s.

5.5 Reliable Data Transfer (RDT)

Ensures packets are delivered and correct.
Uses ACK (acknowledged) and NAK (resend).
TCP implements this via sequence numbers and retransmission.

6. Transport Layer & TCP

6.1 TCP 3-Way Handshake

The 3 steps to establish a reliable TCP connection:

SYN – Client: "Can we talk?"
SYN-ACK – Server: "Yes!"
ACK – Client: "Let’s go!"

After this, both sides can send/receive data.

6.2 Why Use UDP Instead of TCP?

UDP allows more control over:
- What data is sent in each segment.
- When the segment is sent.

6.3 TCP Sequence and Acknowledgment Numbers

SEQ# (Sequence Number): Tells where data fits in the stream.
ACK# (Acknowledgment Number): Confirms receipt and tells what’s expected next.

Formula:

ACK# = SEQ# + data length
Next SEQ# = ACK#

6.4 Cumulative Acknowledgment (ACK)

TCP sends one ACK for multiple received segments.
Saves bandwidth, simplifies recovery, and increases efficiency.

6.5 TCP Window Size

The TCP window limits how much unacknowledged data can be in flight.
Think of it like mailing only 5 letters at once—wait for a reply before sending more.

6.6 Go-Back-N (Sliding Window Protocol)

Sender can send multiple segments up to the window size.
If one fails, resend that one + all after it.

Summary of Go-Back-N Cases:

Case A (Segment 2 corrupted): Retransmit 2–5 → Total 9 segments.
Case B (ACK for 2 lost): No resend needed due to ACK 3 (cumulative).
Case C (ACK for 5 lost): Retransmit segment 5 → Total 6 segments.

6.7 TCP Flow Control

If Receiver's read speed < Sender's speed, TCP flow control reduces sender's rate.
Controlled by rwnd (Receiver Window).
- rwnd = 0: pause sending
- rwnd > 0: resume

7. Switch Fabric & Packet Conversion

7.1 Decimal to Binary Conversion

Method 1 (Subtracting powers of 2): Break the decimal into binary by subtracting the highest powers of 2.

Method 2 (Divide by 2): Divide decimal by 2 repeatedly. Binary is the remainders in reverse order.

Example:

202 → 11001010
3 → 00000011
14 → 00001110
25 → 00011001

Final IP in binary:

11001010.00000011.00001110.00011001

Binary to Decimal: Multiply each bit by powers of 2 from right to left and sum the results.

7.2 Switch Fabric Overview

Types:

Memory-based switching: All packets go through a shared memory. One at a time.
Bus-based switching: Shared bus. Only one packet can use it at once.
Crossbar switching: Allows multiple packets to go through simultaneously if different outputs.

Delay Summary:

Fabric Type	Max Delay
Memory	(N - 1) × D
Bus	(N - 1) × D
Crossbar	0

7.3 Switch Architecture Scheduling

Round Robin (RR)

Every flow gets an equal chance to send packets in a rotating fashion.

Weighted Fair Queuing (WFQ)

Each flow is assigned a weight.
Higher weight = More packet-sending opportunities.
Example ratio for weights 0.5 : 0.25 : 0.25 → Serve packets in 2:1:1 pattern.

7.4 IP Header & Fragmentation

IPv4 Header Basics

Total Length field = Header + Payload
Header = 20 Bytes
Carries Layer 4 payload (e.g., TCP, UDP, ICMP)

IP Fragmentation

If a packet is too large, it's split into fragments.
Fields:
- Identification: Same for all fragments.
- Fragment Offset: Position in the original packet.
- Flags: Indicate more fragments.

Fragments are reassembled at the destination.

8. Network Layer Basics & IP Addressing

8.1 What is a Network & IP Address?

A network is a system that allows devices to communicate.
IP address identifies each device on a network.

8.2 What is a Subnet Mask?

Defines how IP addresses are divided into network and host portions.
Subnet masks help route data efficiently.

8.3 Routing vs Forwarding Tables

Routing Table: Like GPS — determines possible routes.
Forwarding Table: Like a driver — decides the actual next step for each packet.

8.4 Forwarding Table Behavior

Routers use destination IP only — not the source — to determine packet path.
They cannot forward differently based on sender.

8.5 Classful IP Addressing

Legacy method with fixed subnet sizes: Class A, B, C.

8.6 CIDR (Classless Inter-Domain Routing)

Allows flexible subnetting using notation like /24, /16.
Routers pick the longest prefix match (most specific).

8.7 Windows Network Configuration Example

IP Range: 10.1.4.0/24
Gateway: 10.1.4.1

8.8 Longest Prefix Match (LPM)

Routers choose the route with the most matching bits in the network address.

8.9 DHCP (Dynamic Host Configuration Protocol)

Assigns IP automatically using 4 steps (DORA):
- Discover → Offer → Request → Acknowledge
Provides IP, Gateway, DNS.

8.10 NAT (Network Address Translation)

Allows private IPs to access public internet.

Static NAT

One-to-one permanent IP mapping.

Dynamic NAT

Uses a pool of public IPs, assigned temporarily.
Limited by available public IPs.

PAT (Port Address Translation)

Many private IPs share a single public IP.
Uses different port numbers to track sessions.

9. Routing Algorithms

9.1 Dijkstra’s Link State Routing

Goal: Find the shortest path from a source node to all others.
Steps:
1. Start at source, initialize distances (0 for source, ∞ for others).
2. Check direct neighbors, record link costs.
3. Select the node with the lowest tentative cost.
4. Update costs for neighbors of that node.
5. Repeat until all nodes are visited.

Terms:

N′: Set of nodes with known shortest paths.
D(x): Shortest distance from source to node x.
p(x): Previous node on the shortest path to x.

Shortest Path Example (u to z):

Result: u → v → x → z
Total cost = 14

9.2 Bellman-Ford’s Distance Vector Routing

Each router maintains a routing table with cost to each destination.
Routers send their tables to neighbors.
Update table if a shorter path is found.
Repeat until no more updates (network converges).

Formula:

New Cost to Destination = Cost to Neighbor + Neighbor’s Cost to Destination

Problem Summary: Router x:

Connected to:
- y with cost 4 (and y→u = 11)
- w with cost 6 (and w→u = 9)

To update:

If c(x, w) < 5: route via w is shorter → update & notify.
If c(x, y) > 5: route via w becomes better → update.

No update:

If c(x, w) ≥ 5: no better than 14.
If c(x, y) < 4: cost reduces slightly but doesn’t change route.

9.3 Link State vs Distance Vector

Feature	Link State (Dijkstra)	Distance Vector (Bellman-Ford)
Information Shared	Entire network topology	Only to neighbors
Update Trigger	On topology change	On receiving updated vectors
Convergence Speed	Fast	Slower, can have loops
Computation	Local Dijkstra algorithm	Iterative table updates

10. Autonomous Systems

10.1 Border Gateway Protocol (BGP)

BGP connects Autonomous Systems (AS) across the internet.
Chooses best path to send data based on multiple factors, not just shortest AS-path.
Loop prevention: BGP checks for its own AS in AS-PATH and rejects if found.

Example: UTS (AS7575) → Telstra (AS1221) → Google (AS15169).
If AS7575 sees itself in the path again, it drops the route.

10.2 BGP Path Selection and Economics

A longer AS path may be preferred for cost/partnership reasons.
UTS may choose:
- Path 1: Short (2 hops) via ISP-A
- Path 2: Longer (3 hops) via ISP-B (preferred due to cost/relationship)

10.3 BGP vs IGPs (RIP, OSPF)

Between ASes:
- eBGP: External BGP between ASes.
- iBGP: Internal BGP within the same AS.

11. Network

11.1 Parity Check

What is Parity?
A method to detect transmission errors using an extra bit to ensure the number of 1s is even (even parity) or odd (odd parity).

Even Parity Example (2D Parity):
Each row and column must have an even number of 1s.
Single-bit errors can be located and corrected.
Double-bit errors can be detected but not corrected.

11.2 Address Resolution Protocol (ARP)

Definition:
ARP (Address Resolution Protocol) maps an IP address (Layer 3) to a MAC address (Layer 2) within the same LAN.

How it Works:

PC1 sends an ARP request: "Who has IP X?"
Target device responds with its MAC address.
Sender saves it in ARP cache and sends the data.

11.3 IP & MAC Assignment (Network Example)

Subnets:

Subnet 1 → 192.168.1.xxx
Subnet 2 → 192.168.2.xxx
Subnet 3 → 192.168.3.xxx

Each host and router interface is assigned an IP and MAC address. These addresses are essential for routing and ARP resolution.

11.4 ARP Table (Up-to-date)

Host E to Host B:

Host E (in Subnet 3) sees B is in another subnet → sends to default gateway (Router 2).
R2 forwards to R1 → R1 checks destination → sends to Host B via MAC.

MAC changes at each hop, IP stays the same.

11.5 ARP Table (Empty)

Host E to Host B with Empty Table:

Host E sends ARP request: “Who has IP 192.168.3.2?”
Router 2 replies with MAC.
Host E sends real packet using MAC.
Same routing process as before resumes.

11.6 Router Replaced by Switch (Scenario)

a. Host E to Host F

Same subnet → E uses ARP to get F’s MAC → sends directly (no router).

b. Host E to Host B

Different subnet → E sends to default gateway → ARP for Router R1’s MAC.

c. Host A to Host B

Same subnet, no MAC in cache → A sends ARP broadcast.
Switch floods ARP to all ports.

Switch S1 Actions:

Floods the ARP request.
Learns MAC from incoming frames.

Router R1 Action:
Does not forward ARP broadcasts (Layer 2).

Host B Action:
Learns A’s MAC from the ARP request.

Switch After Reply:
Updates MAC table and forwards unicast reply to A.

1. Internet Structure & Network Fundamentals

1.1 Protocol​

1.2 Internet Structure​

Key Components:​

1.3 Access Networks​

DSL (Digital Subscriber Line)​

HFC (Hybrid Fibre-Coaxial)​

NBN (National Broadband Network)​

Enterprise Networks​

1.4 Physical Media​

1.5 Packet Switching & Delays​

Types of Delay:​

1.6 Network Core Functions​

1.7 Circuit Switching​

Multiplexing:​

1.8 Packet Switching vs Circuit Switching​

1.9 Internet Structure​

2. Networking Performance & Internet Stack

2.1 Sources of Packet Delay​

Transmission Delay​

Propagation Delay​

Queueing Delay​

Processing Delay​

2.2 Traffic Intensity (La/R)​

2.3 Packet Loss​

2.4 Throughput​

2.5 Internet Protocol Stack (TCP/IP)​

2.6 Packet Overlap & Queue Test (Visual Method)​

Visual Method​

2.7 TCP/IP Model Summary Table​

3. Application Layer (1 of the 5 Layers)

3.1 What an Application-Layer Protocol Defines​

Protocol Examples​

3.2 Internet Transport Protocol Services​

Example Transport Use by Application​

3.3 Web and HTTP​

What is a Web Page?​

HTTP: Stateless Protocol​

Persistent vs Non-Persistent HTTP​