Comprehensive Study Notes: Data Communications and Computer Networks (R22A0411)

Introduction to Data Communications and System Components

Data communication is the sharing of information across distances, occurring either locally (face-to-face) or remotely. A data communications system is composed of five fundamental components. The message is the information or data to be shared, taking forms such as text, numbers, pictures, audio, or video. The sender is the device that transmits the message, which can include computers, workstations, telephones, or video cameras. The receiver is the destination device, such as a computer or television. The transmission medium is the physical pathway, such as twisted-pair wire, coaxial cable, fiber-optic cable, or radio waves, through which the message travels. Finally, a protocol is a set of rules governing the communication; it represents an agreement between devices to ensure they can understand each other despite differing internal architectures.

Data is represented in various formats within these systems. Text is represented as a bit pattern, a sequence of zeros and ones, through coding systems like Unicode, which uses 32 bits to represent symbols in any language. The first 127 characters of Unicode originate from the American Standard Code for Information Interchange, commonly known as ASCII. Numbers are typically converted directly to binary to simplify mathematical operations. Images are represented by a matrix of pixels, where higher resolution (more pixels) requires more memory. For black-and-white images, a 1-bit pattern per pixel is sufficient, while grayscale and color images (using RGB or YCM methods) require larger bit patterns to signify intensity levels. Audio involves continuous sound signals changed into digital or analog formats, while video can be a continuous entity or a discrete series of images arranged to convey motion.

Data flow between devices occurs in three modes: simplex, half-duplex, and full-duplex. In simplex mode, communication is unidirectional, similar to a one-way street; one device only transmits while the other only receives, as seen with keyboards and monitors. Half-duplex mode allows both stations to transmit and receive, but not simultaneously; the entire channel capacity is used by the transmitting device at any given time, much like a one-lane road. Full-duplex mode allows both stations to transmit and receive simultaneously by sharing the channel capacity through physically separate paths or by dividing the channel bandwidth, as demonstrated by telephone networks.

Network Criteria, Structure, and Topologies

A network is defined as a set of nodes connected by communication links. These nodes can be computers, printers, or other data-capable devices. Most networks employ distributed processing, where tasks are divided among multiple computers rather than a single machine. For a network to be effective, it must meet criteria regarding performance, reliability, and security. Performance is measured by transit time and response time, and it is evaluated by metrics such as throughput and delay. Reliability is measured by the frequency of failure, recovery time from failure, and robustness during catastrophes. Security involves protecting data from unauthorized access or damage.

Physical structures within a network include point-to-point and multipoint connections. A point-to-point connection provides a dedicated link between two devices, while a multipoint connection allows more than two devices to share a single link either spatially (simultaneously) or temporally (taking turns). The physical topology refers to the geometric representation of these links. The mesh topology requires every device to have a dedicated point-to-point link to every other device; in a fully connected mesh with $n$ nodes, the number of duplex-mode links is calculated as $\frac{n \times (n - 1)}{2}$. Each node must have $n - 1$ input/output ports. While robust and secure, mesh topologies are expensive and difficult to install due to the massive volume of cabling required.

In a star topology, each device has a dedicated point-to-point link to a central controller known as a hub. Direct traffic between devices is not allowed; the hub acts as an exchange to relay data. Star networks are less expensive and easier to reconfigure than mesh networks, but they suffer from a single point of failure if the hub goes down. A bus topology is a multipoint connection where one long cable acts as a backbone with nodes connected via drop lines and taps. This topology is easy to install but difficult to reconfigure, and a break in the backbone cable stops all transmission. A ring topology links each device with a point-to-point connection to the two devices on either side, using repeaters to pass signals in one direction. Hybrid topologies combine these structures, such as a star backbone connecting several bus sub-networks.

Categories of Networks and the Internet

Networks are categorized by size, transmission technology, and topology. Local Area Networks (LANs) are privately owned networks within a single building or campus, typically operating at speeds from 10 Mbps to 100 Mbps, with newer versions reaching 10 Gbps. They use technologies like Ethernet (IEEE 802.3) or Token Ring (IEEE 802.5). Metropolitan Area Networks (MANs) cover city-sized areas, often utilizing cable television infrastructures and standards like IEEE 802.16. Wide Area Networks (WANs) span large geographical areas, such as countries or continents, using communication subnets composed of transmission lines and switching elements. Most WANs are store-and-forward or packet-switched networks.

The Internet is a massive collaboration of hundreds of thousands of interconnected networks. Its history began in 1969 with ARPANET, which connected four nodes at UCLA, UCSB, SRI, and the University of Utah using Interface Message Processors (IMPs). Key developers Vint Cerf and Bob Kahn outlined protocols that eventually split into Transmission Control Protocol (TCP) and Internetworking Protocol (IP). Today, the Internet is a hierarchical structure run by private companies. International ISPs connect nations; National ISPs are backbone networks connected via Network Access Points (NAPs) or private peering points with data rates up to 600 Mbps. Regional and Local ISPs provide direct service to end users.

Standards, Protocols, and Layered Tasks

Protocols consist of rules governing data communications, characterized by syntax (data structure/format), semantics (meaning of bit sections), and timing (when and how fast data is sent). Standards are divided into de facto (adopted through widespread use) and de jure (legislated by official bodies). Organizations like the ISO, ITU-T, ANSI, IEEE, and EIA develop these standards. Internet standards undergo a specific maturity process, starting as an Internet draft and potentially becoming a Request for Comment (RFC) after successful testing and peer review.

To manage complex networking tasks, the concept of layers is used, similar to a postal system consisting of a sender, carrier, and receiver. The Open Systems Interconnection (OSI) Reference Model, developed by the ISO, consists of seven layers: Physical, Data Link, Network, Transport, Session, Presentation, and Application. The Physical Layer handles raw bit transmission over channels, defining electrical and mechanical specifications. The Data Link Layer transforms the physical medium into a reliable link by creating frames and managing flow and error control. The Network Layer handles source-to-destination packet delivery and routing. The Transport Layer ensures end-to-end, process-to-process delivery of the entire message. The Session Layer manages dialog control and synchronization points. The Presentation Layer handles data translation, encryption, and compression. Finally, the Application Layer provides user-level protocols like HTTP, FTP, and SMTP.

The TCP/IP Reference Model and Addressing

The TCP/IP model consists of four layers: Host-to-Network, Internet, Transport, and Application. The Internet layer utilizes the Internet Protocol (IP) for packet routing and delivery. The Transport layer includes two main protocols: TCP (a reliable, connection-oriented protocol providing byte streams) and UDP (an unreliable, connectionless protocol for applications favoring speed over accuracy). TCP/IP addressing occurs at four levels. Physical addresses (MAC addresses) are 48-bit (6-byte) identifiers printed on Network Interface Cards, expressed in hexadecimal as $07:01:02:01:2C:4B$. Logical addresses (IP addresses) are 32-bit (IPv4) or 128-bit (IPv6) identifiers that are unique globally. Port addresses are 16-bit identifiers ($0$ to $65,535$) used to distinguish between different processes on a single host. Specific addresses include user-friendly identifiers like URLs or email addresses.

Signals and Transmission Media

Information can be represented as analog or digital signals. Analog signals have infinite intensity levels and are often described by characteristics like peak amplitude, period ($T$), and frequency ($f$), where $f = \frac{1}{T}$. Phase describes the status of the first cycle at time zero, measured in degrees or radians. Wavelength ($\lambda$) is calculated as $\lambda = \frac{c}{f}$, where $c$ is propagation speed ($3 \times 10^8 \, \text{m/s}$ in vacuum). Digital signals have discrete levels (usually 1 and 0); if a signal has $L$ levels, it carries $\log_{2}(L)$ bits per level. Bit rate is the number of bits sent per second (bps). Bandwidth refers to the range of frequencies contained in a composite signal.

Transmission media are categorized as guided or unguided. Guided media include twisted-pair cables (unshielded or shielded copper wires twisted to cancel electromagnetic interference), coaxial cables (central copper core with shielding for high frequency), and fiber-optic cables. Fiber-optic cables transmit signals as light using total internal reflection within a glass or plastic core. Advantages include massive bandwidth ($1600 \, \text{Gbps}$ with WDM) and immunity to interference, though they are more expensive and fragile. Unguided media involve wireless communication through radio waves ($3 \, \text{kHz}$ to $1 \, \text{GHz}$), microwaves ($1 \, \text{GHz}$ to $300 \, \text{GHz}$), and infrared ($300 \, \text{GHz}$ to $400 \, \text{THz}$). Propagation modes include ground (hugging earth), sky (ionospheric reflection), and line-of-sight (direct antenna-to-antenna).

Switching and Network Layer Functions

Switching enables one-to-one communication in large networks without redundant point-to-point links. Circuit-switched networks establish a dedicated physical path between stations, involving three phases: setup, data transfer, and teardown. Packet-switching occurs as datagram networks (connectionless, where packets are routed independently) or virtual-circuit networks (connection-oriented, where all packets follow a pre-established path). The Network Layer handles logical addressing, routing, and congestion control. IPv4 uses 32-bit addresses in dotted-decimal notation (e.g., $132.24.75.9$). Classful addressing divides the space into Classes A, B, C, D, and E, while Classless Inter-Domain Routing (CIDR) uses masks to define flexible block sizes (e.g., $x.y.z.t/n$). Network Address Translation (NAT) allows private internal networks to use one or few global IP addresses, conserving the address space.

Data Link Layer, Ethernet, and MAC Protocols

The Data Link Layer is responsible for node-to-node delivery, framing, and error control. Framing techniques include character count, bit stuffing (inserting a $0$ after five consecutive $1$s), and byte stuffing. Error detection involves the Cyclic Redundancy Check (CRC), which uses binary division and generating polynomials, or Checksums, which use one’s complement addition. Multiple access protocols manage shared channels. Random access includes ALOHA (Pure and Slotted) and CSMA (Carrier Sense Multiple Access). CSMA/CD is used in wired Ethernet to detect collisions by monitoring power levels, while CSMA/CA is used in wireless networks (IEEE 802.11) to avoid collisions using interframe spacing (IFS) and Network Allocation Vectors (NAV).

Standard Ethernet (10 Mbps) uses Manchester encoding and a 64-byte minimum frame size to ensure collision detection. Fast Ethernet (100 Mbps) evolved to use star topologies and encoding schemes like 4B/5B. Gigabit Ethernet (1000 Mbps) supports full-duplex operation and carrier extension or frame bursting in half-duplex. Wireless LANs (802.11) use Basic Service Sets (BSS) and Extended Service Sets (ESS). Bluetooth is a short-range wireless technology creating piconets (up to 8 stations) and scatternets using frequency-hopping spread spectrum (FHSS) at 1600 hops per second.

Transport Layer Services: TCP, UDP, and SCTP

The Transport Layer provides process-to-process delivery using port numbers. UDP is a connectionless, unreliable protocol with an 8-byte header (Source Port, Destination Port, Length, Checksum). It is ideal for low-latency applications like gaming and VoIP. TCP is a reliable, connection-oriented protocol using a sequence-numbering system and a 20-byte minimum header. It employs a three-way handshake ($SYN \rightarrow SYN+ACK \rightarrow ACK$) for connection establishment and a four-step process for termination. Error control in TCP is achieved through retransmission based on timers and duplicate ACKs (Fast Retransmit). Flow control uses a sliding window mechanism where the receiver advertises its available buffer space. SCTP (Stream Control Transmission Protocol) is a message-oriented protocol that supports multihoming and multiple streams within a single association, preventing head-of-line blocking while ensuring reliable delivery.

Application Layer Protocols and Network Management

The Application Layer hosts programs like DNS, SMTP, and HTTP. The Domain Name System (DNS) maps domain names to IP addresses through a hierarchical database of root, primary, and secondary servers. Email uses the Simple Mail Transfer Protocol (SMTP) for sending and Post Office Protocol (POP3) or Internet Mail Access Protocol (IMAP) for retrieval. HTTP is the foundation of the World Wide Web, using Request and Response messages (e.g., GET, POST) on port 80. Simple Network Management Protocol (SNMP) allows managers to monitor and control agents (routers/hosts) using the Structure of Management Information (SMI) and Management Information Base (MIB). SNMP PDUs include GetRequest, SetRequest, and Trap messages. Quality of Service (QoS) involves parameters like reliability, delay, jitter, and bandwidth, managed through models like Integrated Services (IntServ/RSVP) or Differentiated Services (DiffServ/DSCP).