Computer Networking, Internet Architecture, and Communication Protocols Study Guide

Fundamentals of Computer Networks and the Internet

Definitions of Core System Terms:
- Net / Network: A system of interconnected computers, devices, and servers that share resources and data.
- Internet: A global network of interconnected computers and devices that communicate with each other using standardized protocols.
- Computing Device: A physical machine capable of running a program (e.g., computers, tablets, servers, routers, smart sensors).
- Computing System: A group of computing devices and programs working together for a common purpose.
- Computing Network: A group of interconnected computing devices capable of sending or receiving data.
- Path: The series of direct connections between computing devices on a network, starting with a sender and ending with a receiver.
- Protocol: An agreed-upon set of rules that specify the behavior of a system.
Structural Complexity and Topologies:
- 2-Device Network: Requires only a physical connection and a shared data format.
- 3-Device Network: Requires an addressing scheme so that devices can distinguish which data is intended for them.
- 6+ Device Topologies:
- Ring Topology: Each device connects to two neighboring devices, forming a circular ring.
- Mesh Topology: Every device is connected directly to every other device on the network.
- Star Topology: One central node/device connects directly to all other devices.
- Bus Topology: All devices connect to a single shared central transmission line.
- Tree Topology: A branching, hierarchical layout combining star and bus elements.
Geographic Network Classification:
- LAN (Local Area Network): Covers a small geographical area, such as a home, school, or office building.
- WAN (Wide Area Network): Spans large geographical regions (cities, countries, globe); composed of multiple interconnected LANs connected via leased telecommunication lines.
- DCN (Data Center Network): Specialized high-performance network used inside data centers where minimal latency and ultra-fast inter-server communication are required.

Physical Infrastructure and Media Types

Physical Transmission Mediums:
- Twisted Pair Copper Cables (Ethernet / CAT5):
- Originates from landline telephone copper infrastructure.
- Contains four twisted pairs (eight individual wires) of copper wire.
- Transmits binary data ( $1$ s and $0$ s) using pulses of electricity (voltage or current).
- Operates under standard Ethernet specifications.
- Maximum Transmission Distance: $330\,\text{ft}$ .
- Bandwidth Capacity: $1\,\text{Gbps}$ .
- Vulnerabilities: Susceptible to electromagnetic interference.
- Applications: Common in both local LANs and national WANs.

Fiber-Optic Cables:
- Contains optical fibers that carry pulses of light (intensity of light) rather than electrical signals.
- Coated with protective plastic layers and encased in a durable outer sheath.
- Follows Ethernet standards for data formatting.
- Maximum Transmission Distance: $50\,\text{miles}$ per segment without amplification.
- Bandwidth Capacity: Up to $26\,\text{Tbps}$ (measured in early 2020).
- Primary Use: Transoceanic lines connecting continents across oceans, as well as high-speed city networks.
- Financial Costs: Tonga-to-Fiji undersea cable cost approximately $\$30\,\text{million}$ ; Google FASTER cable (US to Tokyo) cost $\$300\,\text{million}$ .

Wireless Connections (Wi-Fi / Radio):
- Eliminates initial physical cabling; internal wireless cards convert binary data into electromagnetic radio waves sent through the air.
- Operational Distance: $75\text{--}100\,\text{ft}$ inside obstacle-heavy office environments; up to $1000\,\text{ft}$ in open fields.
- Bandwidth Capacity: Theoretical maximum $1.3\,\text{Gbps}$ (slower in practical conditions).
- Wireless Access Points (WAPs): Intercept radio waves, convert them back to binary data, and transfer them onto wired copper/fiber infrastructure.

Summary Comparison Table of Network Media:
- Wireless | Medium: Radio | Max Distance: $100\,\text{ft}$ | Bandwidth: $1.3\,\text{Gbps}$ | Notes: Slower in real-world conditions.
- Twisted Pair Copper | Medium: Electricity | Max Distance: $330\,\text{ft}$ | Bandwidth: $1\,\text{Gbps}$ | Notes: Susceptible to interference.
- Fiber-Optic Cable | Medium: Light | Max Distance: $50\,\text{miles}$ | Bandwidth: $26\,\text{Tbps}$ | Notes: Expensive installation cost.

Data Transmission Metrics: Bit Rate, Bandwidth, and Latency

Signal Encoding (Line Coding):
- All network communication relies on binary streams ( $1$ s and $0$ s).
- Line Coding: The technique of converting binary data into time-based electromagnetic or optical signals.
- Example: Transmitting decimal $5$ ( $101_2$ ):
- Multi-wire Internal Transfer: Uses 3 wires concurrently (Wire 1 ON, Wire 2 OFF, Wire 3 ON).
- Single-wire Serial Transfer: Transmits over 3 consecutive time intervals over 1 line (Interval 1: ON pulse + wait, Interval 2: OFF/nothing + wait, Interval 3: ON pulse + wait).
Bit Rate Units and Scales:
- Bit Rate: The speed at which data is transmitted, measured in bits per second (bps).
- Historical Bit Rate: Early Internet modems operated at $75\,\text{bps}$ .
- Kilobit (Kbps): $1000^1 = 10^3\,\text{bits}$ ( $1,000\,\text{bits}$ ).
- Megabit (Mbps): $1000^2 = 10^6\,\text{bits}$ ( $1,000,000\,\text{bits}$ ). A $10\,\text{Mbps}$ connection transfers one bit every $100\,\text{ns}$ ( $0.0000001\,\text{s}$ ).
- Gigabit (Gbps): $1000^3 = 10^9\,\text{bits}$ ( $1,000,000,000\,\text{bits}$ ).
- Terabit (Tbps): $1000^4 = 10^{12}\,\text{bits}$ ( $1,000,000,000,000\,\text{bits}$ ).
- Petabit (Pbps): $1000^5 = 10^{15}\,\text{bits}$ ( $1,000,000,000,000,000\,\text{bits}$ ).
Bandwidth Thresholds:
- Bandwidth: The theoretical maximum bit rate of a transmission system.
- Broadband Internet: Standard definition requiring a minimum bandwidth of $256\,\text{Kbps}$ .
- Global Broadband Access: As of 2016, only $40\%$ of residents in developing nations had access to broadband Internet.
Latency and Physical Constraints:
- Latency: The time delay between sending a message and receiving it, measured in milliseconds (ms).
- Round-Trip Latency Calculation: $\text{Round-Trip Latency} = \text{Outbound Travel Time} + \text{Inbound ACK Travel Time}$
- Example: Request travel time = $30\,\text{ms}$ , Response ACK travel time = $40\,\text{ms}$ . Total round-trip latency = $70\,\text{ms}$ .
- Speed of Light Limitation: Electromagnetic and light signals travel at approximately $1\,\text{ft/ns}$ . The physical distance between Los Angeles and Tokyo requires a absolute theoretical minimum latency of $30\,\text{ms}$ .
Network Speed Measurement:
- Network speed depends on both bandwidth and latency.
- Modern connections are often asymmetric: Download speeds (e.g., $39.09\,\text{Mbps}$ ) are higher than upload speeds (e.g., $5.85\,\text{Mbps}$ ) because user activities are dominated by downloading content.

Internet Protocol (IP) and Addressing Hierarchy

IP Address Functionality:
- Every device on the Internet requires a unique IP address to send and receive messages.
- Sender IP addresses are included in headers so destination systems can reply.
IPv4 vs. IPv6 Architecture:
- IPv4 Architecture:
- 32-bit binary address written as 4 decimal numbers separated by dots (octets).
- Octet range: $0$ to $255$ ( $2^8 = 256$ values per octet).
- Total IPv4 Address Space: $2^{32} = 4,294,967,296\,\text{addresses}$
- Representation Example: 74.125.20.113.
- IPv6 Architecture:
- 128-bit binary address represented as 8 hexadecimal blocks separated by colons.
- Block range: 0000 to FFFF ( $2^{16} = 65,536$ values per block, where FFFF equals $65,535$ in decimal).
- Hexadecimal Block Calculation: $\text{FFFF}_{16} = (15 \times 16^3) + (15 \times 16^2) + (15 \times 16^1) + (15 \times 16^0) = 65,535$ $2^{16} - 1 = 65,535$
- Total IPv6 Address Space: $2^{128} = 340,282,366,920,938,000,000,000,000,000,000,000,000\,\text{addresses}$ (340 undecillion addresses).
- Representation Example: 2001:0db8:0000:0042:0000:8a2e:0370:7334.
Dynamic vs. Static Addressing:
- Dynamic IP: Assigned temporarily by Internet Service Providers (ISPs) using an available pool; changes per session or per network location.
- Static IP: Permanently fixed IP address assigned to servers to maintain reachable network locations.
IP Address Hierarchy and Subnetting:
- Structure: Network ID (first sequence of bits) + Host/Node ID (final sequence of bits).
- ISP Hierarchy Example (24.147.242.217):
- First 16 bits (24.147): Comcast network.
- Final 16 bits (242.217): Individual home computer.
- Departmental Subnet Example (141.213.127.13):
- 141.213: University of Michigan network.
- 127: UMich Medicine department subnetwork.
- 13: Specific laboratory computer.
- Subnetting at Bit Level (10001101 11010101 01111111 00001101):
- First 16 bits: University network.
- Next 2 bits: Identifies $2^2 = 4$ distinct departments.
- Final 14 bits: Identifies up to $2^{14} = 16,384$ individual host computers per department.

Network Redundancy and Fault Tolerance

Principles of Redundancy:
- Redundancy: Adding extra communication paths and hardware components so data can traverse alternative routes if a link fails.
- Oakland to New York Routing Paths:
- Path 1 (Direct/Short): Oakland -> Austin -> New York.
- Path 2 (Alternative): Oakland -> Austin -> Tampa -> New York.
Fault Tolerance and Single Points of Failure:
- Fault-Tolerant System: A system designed to continue functioning without interruption despite individual component failures.
- Single Point of Failure: A component whose individual failure causes the complete breakdown of the entire system.
Subsea Cable Real-World Case Studies:
- High Redundancy Regions: The North Atlantic Ocean has dozens of submarine cables connecting North America and Europe, eliminating single points of failure across the ocean.

Low Redundancy Regions: South Pacific islands rely on minimal physical cable links.

Tonga Cable Cut (2019): An anchor drag severed the single undersea fiber cable connecting Tonga to Fiji, disrupting national Internet access for 11 days. Satellite internet was brought in as low-bandwidth backup (graceful degradation).
Georgia Cable Cut (2011): An elderly woman digging for scrap metal cut a landline fiber cable with a shovel, causing Armenia to lose Internet access for 5 hours.
Global Cable Cut Rate: Undersea fiber optic cables experience accidental cuts approximately once every 3 days worldwide.

Transport Layer Protocols: TCP vs. UDP

Packet Transmission Fundamentals:
- Datastream: Information broken down into smaller chunks for transmission over IP networks.
- Packet: A discrete chunk of data transmitted across a network.
- Packet Metadata: Additional header information attached to packets (source/destination IP addresses, sequence numbers, checksums) used for routing and message reassembly.
User Datagram Protocol (UDP):
- Lightweight, connectionless transport protocol operating on top of IP.
- Features: Fast, low-latency, minimal overhead. Does not perform retransmission of dropped packets, reordering, or connection handshakes (often termed an Unreliable Data Protocol).
- Target Use Cases: Real-time applications where speed is prioritized over losslessness (e.g., live video streaming, voice calls, multi-player online gaming, DNS queries).
- UDP Segment Layout:
- Header Size: Fixed 8 bytes ( $64\,\text{bits}$ ).
- Source Port (2 bytes / 16 bits) & Destination Port (2 bytes / 16 bits): Port numbers direct traffic to virtual endpoints on a device.

- **Segment Length** (2 bytes / 16 bits): Stores total byte length of segment (header + data). Maximum theoretical segment size:

$2^{16} - 1 = 65,535\,\text{bytes}$ - Checksum (2 bytes / 16 bits): Field used to verify packet data integrity.

Step-by-Step UDP Checksum Verification:
- Sender Algorithm:
1. Encodes text string "Hola" into binary using ASCII/UTF-8:
  - 'H' = 01001000
  - 'o' = 01101111
  - 'l' = 01101100
  - 'a' = 01100001
2. Groups bytes into 16-bit blocks: 0100100001101111 and 0110000101101100.
3. Sums 16-bit blocks: $\begin{aligned} 0100100001101111_2 &(18,543_{10}) \\ + 0110110001100001_2 &(27,745_{10}) \\ \hline 1011010011010000_2 &(46,288_{10}) \end{aligned}$
4. Inserts 1011010011010000 into Checksum field.
- Receiver Detection of Corrupted Packet ("Mola"):
1. Corrupted 'M' binary: 01001101.
2. Grouped 16-bit blocks: 0100110101101111 and 0110110001100001.
3. Recalculates sum: $\begin{aligned} 0100110101101111_2 &(19,823_{10}) \\ + 0110110001100001_2 &(27,745_{10}) \\ \hline 1011100111010000_2 &(47,568_{10}) \end{aligned}$
4. Compares calculated checksum (1011100111010000) with packet checksum (1011010011010000). Because values differ, packet corruption is confirmed and discarded.
Transmission Control Protocol (TCP):
- Connection-oriented, reliable transport protocol operating on top of IP (TCP/IP stack).
- Header Size: Variable between 20 bytes and 60 bytes.
- Features: Guarantees packet order, data integrity, and delivery via acknowledgments and retransmissions.
- Protocol Lifecycle:
1. - Establish Connection (3-Way Handshake):
    - Step 1: Client sends TCP packet with SYN flag set to $1$ (SYN).
    - Step 2: Server responds with packet containing SYN = $1$ and ACK = $1$ (SYN-ACK).
    - Step 3: Client replies with packet containing ACK flag set to $1$ (ACK).
  - Data Transmission and Tracking:
    - Each segment carries a 32-bit Sequence Number and Acknowledgement Number.
    - Receiver increments Acknowledgement Number by total payload bytes received.
    - Timeout Retransmission: Sender places outgoing packets in a timer queue. If timer expires before receiving an ACK, sender retransmits packet.
    - Handling Out-of-Order Packets: Receiver uses sequence numbers to buffer and resequence packets into original order.
  - Terminating Connection:
    - Initiating host sends packet with FIN flag set to $1$ (FIN).
    - Remote host responds with ACK followed by FIN.
    - Initiating host confirms with final ACK to close socket.

The Domain Name System (DNS)

Functionality:
- Acts as a distributed phonebook translating human-readable domain names into IP addresses.
Domain Anatomy:
- Structure: [third-level-domain].[second-level-domain].[top-level-domain]
- Top-Level Domain (TLD): Managed global endings such as .com, .org, .edu.
- Second-Level Domain: Specific organizational name registered under TLD (e.g., wikipedia, khanacademy).
- Third-Level Domain (Subdomain): Specific server partition assigned by domain owner (e.g., www, es, m).
Resolving a Domain Name (DNS Resolution Steps):
1. Local Browser Cache: Checks locally stored mapping history (chrome://net-internals/#dns).
2. ISP Resolver Cache: Queries Internet Service Provider's shared lookup cache.
3. Root Name Servers: ISP queries root servers to direct to the appropriate TLD server.
4. TLD Name Servers: Directs ISP to host name servers managing specific domain records.
5. Host Name Servers: Returns exact authoritative IP address to ISP and user device.

The World Wide Web and HTTP

Distinction between Web and Internet:
- Internet: The underlying physical network infrastructure and core IP protocols.
- World Wide Web (WWW): An application network of HTML documents, media, and files linked via URLs running on top of the Internet.

HTTP Transmission Sequence:

User Directs Browser to URL: Enters address starting with http:// or https://.
DNS Mapping: Resolves host domain to numeric IP.
Browser Transmits HTTP Request

text HTTP/1.1 200 OK Content-Type: text/html; charset=UTF-8 Content-Length: 208

… ```

HTTP Status Codes: 200 OK (Successful request), 404 File Not Found (Missing document error).

Rendering Content: Browser interprets HTML markup and executes external scripts/styles.

System Scalability and Load Testing

Scalability Factors:
- Scalable System: A system that maintains performance without crashing as load or user volume increases.
- Network Drivers: Open protocol standards, dynamic IP routing, and 128-bit IPv6 address spaces support network scaling.
Physical Scalability Bottlenecks:
- Local access link bandwidth limitations.
- Router Throughput: Consumer routers handle $\,\text{1 Gbps}$ throughput; enterprise routers handle up to $10\,\text{Gbps}$ .
- Wi-Fi Access Point Device Limits: Routers typically cap simultaneous connections around $250\,\text{devices}$ .
Real-World Load Spikes:
- Khan Academy Traffic Spike (March 2020): Remote learning due to COVID-19 caused a $250\%$ load increase. The student assignment notification queue accumulated a massive task backlog until engineers increased processing capacity.

Pokémon Go Launch (Summer 2016): Engineers performed load testing at $5\times$ estimated worst-case traffic. On launch day, actual traffic reached $50\times$ estimates, overwhelming Cloud Datastore transaction limits.

Internet Governance, Protocols, and Social Dilemmas

Standards Organizations & Protocol Classifications:
- IETF (Internet Engineering Task Force): Open international organization that develops and maintains open Internet standards and protocols (TCP/IP, HTTP, DNS).
- Open Protocols: Non-proprietary standards accessible to all developers and manufacturers, ensuring full device interoperability.
- Proprietary Protocols: Controlled exclusively by a single company, limiting communication between devices of different vendor ecosystems.
Internet Policy Dilemmas:
- Net Neutrality: Principle that Internet service providers must treat all data equally, without favoring, throttling, blocking, or charging extra for specific services or websites.
- Internet Censorship: Control or suppression of published or accessed content on the Internet by authorities or institutions. Used for legal protection, but can also limit free expression.