Overview of Computer Networks and Data Transmission

This comprehensive guide focuses on the fundamental concepts, definitions, and architectures within the realm of computer networks and data transmission. The curriculum aims to provide a deep understanding of the communication models, network classifications, and layered architectures, specifically referencing the Open Systems Interconnection (OSI) and Transmission Control Protocol/Internet Protocol (TCP/IP) models.

Fundamental Concepts and Definitions of Computer Networks

A computer network (Computer Network) is defined as a system consisting of individual computers, known as network nodes, that are interconnected through a physical transmission medium (physical media). These nodes follow a specific network architecture and possess the capability to exchange information. The physical transmission media can be categorized into two primary types: wired (hữu tuyến) and wireless (vô tuyến). Wired media include twisted-pair cables, fiber optic cables, and coaxial cables. Wireless media encompass infrared waves, microwave (viba), and satellite communications.

Network architecture refers to the specific methodology used for connecting nodes and facilitating information exchange. It consists of two main pillars: Topology and Protocol. Network topology (Hình trạng mạng) describes the geometric arrangement or physical layout of how nodes are connected. Network protocol (Giao thức mạng) represents the set of rules that dictate how nodes exchange information with one another.

Physical and Logical Network Topologies

The Star Topology (Mạng dạng sao) features a central device that coordinates all activities within the system. This central hub identifies the pairs of sender and receiver addresses authorized to occupy the communication line, approves and monitors the exchange process, and broadcasts status messages regarding the local area network (LAN). Its advantages include continued operation even if a single node fails, a simple structure that allows for stable control algorithms, and the flexibility to expand or shrink based on user needs. However, if the central device fails, the entire system collapses. Connecting distances are typically limited to approximately $100\,m$ , and expansion is strictly dependent on the capacity of the central device.

The Bus Topology (Mạng dạng trục) uses a single main cable with both ends sealed by terminators. This setup is cost-effective regarding cable length and is relatively easy to install. However, it is prone to traffic congestion during high-volume data transmissions. Identifying faults at specific segments is difficult, and a break in the main line can halt the entire system.

The Ring Topology (Mạng dạng vòng) is arranged in a circular fashion where the cable forms a closed loop, and signals travel in one specific direction. While it can cover long distances with less cabling than other types and offers faster speeds than the Bus topology, it is rarely used in practice. Its weaknesses involve slow overall performance compared to modern standards, system-wide failure if any part of the cable is damaged, and difficulty in troubleshooting errors.

The Mesh Topology (Mạng dạng lưới) is utilized for mission-critical systems where downtime is unacceptable, such as nuclear power plants or national security networks. In this structure, every computer is connected to every other computer in the network. This comprehensive connectivity is the foundational structure of the global Internet.

The Extended Star topology is a combination of multiple star networks linked through hubs or switches. This allows for the expansion of the physical size and distance of a standard star configuration. Overall, the benefits of networking include the sharing of resources (printers, hard drives) regardless of geographical location, standardized security and data safety across multiple terminals, high system reliability via component replacement, and the creation of a global communication environment without boundaries.

Network Components and Host Roles

Every computer on a network is referred to as a host or an end device. Within the network, devices take on specific roles. Servers are computers that provide information to other devices. Examples include Email servers running specialized software to manage electronic mail, Web servers running software to host and serve web pages, and File servers which store and manage collaborative files. Clients are computers that send requests to these servers to retrieve information, such as accessing a website via a browser or downloading an email.

In a Peer-to-Peer (P2P) network, a device can act as both a client and a server simultaneously. This design is recommended only for very small networks. For instance, one PC might have a Universal Serial Bus (USB) connection to a printer and a network interface card (NIC) connection to another PC for file sharing. P2P networks are easy to set up, less complex, and low cost, but they suffer from decentralized management, lack of security, poor scalability, and lower performance.

Intermediary Network Devices connect end devices and manage data as it traverses the network. These include switches, wireless access points, routers, and firewalls. Their roles include regenerating and retransmitting data signals, maintaining information about available network paths, and notifying other devices of communication errors or failures.

Network Media provides the channel through which messages travel from source to destination. Modern networks use three media types: metallic wires within cables using electrical pulses, glass or plastic fibers (fiber optics) using light pulses, and wireless transmission using specific frequency modulation of electromagnetic waves.

Principles of Communication Systems and Network Architecture

The fundamental purpose of a communication system is the exchange of data between two entities. A typical model involves a source system (such as a workstation connected to a modem), a transmission system (like a public telephone network), and a destination system (a modem connected to a server). When designing network architecture, four basic characteristics must be addressed: Fault Tolerance, Scalability, Quality of Service (QoS), and Network Security.

Fault Tolerance limits the impact of a failure by ensuring that many paths are available for data transmission. Reliable networks use packet switching, which divides traffic into packets that can each take a different path to the destination. This is impossible in circuit-switching networks that establish dedicated circuits. Scalability ensures the network can grow quickly to support new users and applications without degrading performance for existing users, achieved through the use of accepted standards and protocols.

Quality of Service (QoS) is vital for live voice and video, which require high expectations for delivery. Without QoS, high bandwidth demand leads to interruptions. Routers use QoS mechanisms to prioritize traffic; for example, a Voice over IP (VoIP) call is given priority over a standard web page download to maintain a smooth experience. Network Security involves two main areas: Infrastructure security (physical protection and preventing unauthorized access) and Information security (protecting the data transmitted). The three goals of security are Confidentiality (only intended recipients can read data), Integrity (ensuring data is not altered), and Availability (reliable access for authorized users).

Network Representation and Protocols

Network representations utilize topology diagrams with specific symbols for devices. Key terms include the Network Interface Card (NIC), Physical Port, and Interface. Physical topology diagrams illustrate the physical location of devices and cables, while Logical topology diagrams map out the devices, ports, and addressing schemes. Logical topologies can be Point-to-Point (a dedicated physical channel between two nodes) or Multi-point/Broadcast (all nodes share a common medium).

Point-to-Point logic often results in low information collision but suffers from low transmission efficiency, high resource consumption, and high latency due to the signaling time at intermediate nodes. This is also called a Store-and-Forward network. In Broadcast logic, every node receives a message and checks if the destination address matches its own. This requires collision or congestion resolution mechanisms like those seen in Bus and Ring networks.

Protocols are sets of rules for communication. Every communication requires a message source, a message destination, and a channel. Computer protocols must handle message encoding (converting info into a transmittable form), formatting and encapsulation (specific structures for specific channels), message size, timing, and delivery options. Message timing includes Flow Control (managing the rate of data), Response Timeout (actions taken when no reply is received), and Access Methods (determining when a device can send to avoid collisions). Delivery options include Unicast (to one device), Multicast (to a group), and Broadcast (to all devices).

History and Development of the Internet

The development of computer networks began in $1961$ with Kleinrock’s queuing theory proving the efficiency of packet switching. In $1964$ , Baran proposed packet switching for military use. By $1969$ , the US Department of Defense built ARPANET, the precursor to the Internet. In $1970$ , the ALOHAnet satellite network launched in Hawaii. $1972$ saw the public announcement of ARPANET, the first host-host protocol (NCP), and Ray Tomlinson’s invention of E-mail when ARPANET had only $15$ nodes.

In $1973$ , Metcalfe’s doctoral thesis proposed Ethernet. $1974$ was a landmark year as Vinton Cerf and Bob Kahn created the TCP/IP protocol, and the term "Internet" was first used. High-level services like FTP were developed by AT&T in $1976$ . By $1982$ , TCP/IP became the standard protocol for ARPANET. The network split in $1983$ , with part becoming NSFnet. By the late $1990s$ , the web was commercialized, P2P and instant messaging emerged, and backbone speeds reached the Gbps range for over $100\,\text{million}$ users.

In Vietnam, the first connection attempts occurred in $1991$ . By $1996$ , VNPT launched ISP services at $64\,kbps$ . On November $19$ , $1997$ , Vietnam officially connected to the global Internet. By $2007$ , there were $20$ ISPs and $19\,\text{million}$ users. Key entities include IAPs (Internet Exchange Providers like VNPT, FPT, Viettel), ISPs (Internet Service Providers), ICPs (Content Providers), and OSPs (Online Service Providers).

Network Classification and Metrics

Networks are classified by scale or transmission technique. By scale: Local Area Network (LAN) covers a building or campus (up to a few $km$ ); Metropolitan Area Network (MAN) covers a city (up to $100\,km$ ); Wide Area Network (WAN) covers countries/continents (up to $1000\,km$ ); and Global Area Network (GAN) covers the world.

Transmission performance is measured by Bandwidth ( $R$ ) in bits per second (bps) and Latency. The mathematical models for delay are as follows:

$\text{Trễ truyền tải} = \frac{\text{Kích thước dữ liệu}}{\text{Băng thông}}$

$\text{Trễ truyền dẫn} = \frac{\text{Độ dài liên kết}}{\text{Tốc độ tín hiệu}}$

The signal speed is approximately $2 \times 10^8\,m/s$ .

Transmission techniques include Circuit Switching, where resources are dedicated to a connection through three phases: Circuit establishment, Data transmission, and Disconnect. This ensures QoS for real-time audio/video but is inefficient when the line is idle. Message Switching (Store-and-Forward) treats messages as independent blocks and does not use dedicated links, improving efficiency but increasing latency. Packet Switching divides data into small packets with headers and trailers. Packets are routed independently, can arrive out of order, or be lost, requiring the receiver to reconstruct the data.

Layered Architecture Models

Complex network operations are explained using layered models to modularize activities. Each layer provides services to the layer above and consumes services from the layer below. Data is processed into Protocol Data Units (PDU), consisting of a Header and a Payload. Communication is either Connection-oriented (Reliable, high overhead, $3$ phases) or Connectionless (Unreliable, "best effort," no setup).

The OSI Model, established by the International Organization for Standardization (ISO) in $1984$ , has seven layers:

Physical: Transmission of unstructured bits.
Data Link: Reliable transfer of frames.
Network: Path determination and logical addressing.
Transport: End-to-end reliability and data segmentation.
Session: Managing session dialogs.
Presentation: Data representation, encryption, and compression.
Application: User-network interface.

The TCP/IP Model is the practical standard used today and consists of four layers: Application, Transport, Internet, and Network Access. Encapsulation occurs as data moves down the stack (adding headers/trailers), and De-encapsulation occurs as data moves up the stack at the receiving end. While the OSI model is a theoretical reference, TCP/IP is the implementation used across almost all network systems.

Key Abbreviations and Terminology

GAN: Global Area Network (Mạng toàn cầu).
IP: Internet Protocol (Giao thức liên mạng).
ISO: International Organization for Standardization (Tổ chức tiêu chuẩn quốc tế).
LAN: Local Area Network (Mạng cục bộ).
MAN: Metropolitan Area Network (Mạng đô thị).
MTU: Maximum Transfer Unit (Độ dài gói tin cực đại).
OSI: Open System Interconnection (Mô hình kết nối các hệ thống mở).
PDU: Protocol Data Unit (Đơn vị dữ liệu giao thức).
TCP: Transmission Control Protocol (Giao thức điều khiển truyền dẫn).
VC: Virtual Circuit (Kênh ảo).
WAN: Wide Area Network (Mạng diện rộng).