Network System Computing and Network-Designing
Network computing is the foundational concept of linking independent computing devices to share resources, data, and processing power. Instead of a single computer operating in isolation, systems work collectively over a shared medium.
Centralized Computing: A single powerful central host (e.g., a mainframe) handles all processing, data storage, and control. Connected clients act as "dumb terminals" that merely handle input and output.
Distributed Computing: Processing tasks, data, and applications are physically separated across multiple independent computers (nodes) connected via a network. They coordinate their actions by passing messages.
Client-Server Model: A distributed structure that partitions workloads between service providers (servers) and service requesters (clients). The client initiates a request, and the server processes it and returns the output.
Peer-to-Peer (P2P) Model: A decentralized model where every workstation (peer) has equivalent capabilities and responsibilities. There is no centralized server; each node acts as both a client and a server to share files or processing tasks directly.
Bandwidth: The maximum theoretical capacity of a communication channel to transmit data over a specific period, typically measured in bits per second (bps, Mbps, Gbps).
Throughput: The actual volume of data successfully transmitted over the network in a given timeframe under real-world conditions (accounting for overhead, errors, and congestion).
Latency: The total time delay incurred for data to travel from its source node to its destination node across the network.
Jitter: The statistical variance or fluctuation in packet arrival times, critical for real-time traffic like voice and video.
Packet Loss: The percentage of transmitted data packets that fail to reach their destination due to network congestion, faulty hardware, or signal degradation.
Network Architecture defines the conceptual design, operational frameworks, and standardized structural configurations used to establish communication across nodes. It dictates the rules, protocols, and logical organization of data transmission.
An abstract, theoretical framework defining network functions into seven distinct logical layers. Each layer serves the layer above it and receives services from the layer below it.
+--------------------------------------------------------------------------------------------+
| Layer 7: Application | -> End-user protocols (HTTP, FTP, SMTP)
+---------------------------------------------------------------------------------------------+
| Layer 6: Presentation | -> Data formatting, encryption, syntax
+----------------------------------------------------------------------------------------------+
| Layer 5: Session | -> Managing communication sessions
+----------------------------------------------------------------------------------------------+
| Layer 4: Transport | -> End-to-end connections, flow control (TCP, UDP)
+-----------------------------------------------------------------------------------------------+
| Layer 3: Network | -> Logical addressing, routing packets (IP)
+------------------------------------------------------------------------------------------------+
| Layer 2: Data Link | -> Physical addressing, framing, MAC/LLC (Ethernet)
+-------------------------------------------------------------------------------------------------+
| Layer 1: Physical | -> Electrical, optical, or radio signals
+-------------------------------------------------------------------------------------------------+
The real-world, functional architecture upon which the internet and modern networks operate. It condenses the theoretical OSI model into four practical layers:
Application Layer: Combines OSI Layers 5, 6, and 7. Contains high-level protocols (HTTP, HTTPS, SSH, DNS, BGP).
Transport Layer: Maps directly to OSI Layer 4. Manages host-to-host communication via connection-oriented transmission (TCP) or connectionless transmission (UDP).
Internet Layer: Maps to OSI Layer 3. Handles packet construction, logical addressing, and routing across networks using IPv4 or IPv6.
Network Access Layer: Combines OSI Layers 1 and 2. Defines how data is physically framed and transmitted over the media (Ethernet, Wi-Fi, Fiber optics).
Topologies describe how devices are logically or physically arranged and how data flows through them.
Bus Topology: All nodes connect to a single central coaxial cable (the bus). Data travels linearly. A break in the main cable takes down the entire network. Requires termination at both ends.
Star Topology: Every individual node connects directly to a central hub, switch, or router. If a peripheral cable fails, only that node is isolated. If the central device fails, the whole network collapses.
Ring Topology: Each device connects exactly to two neighboring devices, forming a continuous circular loop. Data travels in one direction (or two in a counter-rotating ring) via token passing.
Mesh Topology: Every node connects to multiple other nodes.
Full Mesh: Every single node is interconnected with every other node, offering maximum redundancy but high deployment costs.
Partial Mesh: Select vital nodes are interconnected, while less critical nodes connect only to specific endpoints.
Tree (Hierarchical) Topology: A combination of star and bus topologies, arranged in groups of star-configured networks connected to a central linear backbone bus.
Network Infrastructure represents the physical, tangible hardware, transmission media, and base software configurations that instantiate the logical architecture.
Network Interface Card (NIC): Hardware component built into a device that allows it to connect to a network. It contains a hardcoded, unique 48-bit physical address known as the MAC Address.
Repeater: A Layer 1 device that receives electrical, optical, or radio signals and regenerates them to extend the physical range of the transmission medium.
Hub: A legacy, non-intelligent Layer 1 multiport device that broadcasts incoming data packets out of all its ports to every connected device, creating a single large collision domain.
Bridge: A Layer 2 device that filters and forwards traffic between two distinct network segments based on MAC addresses, splitting a collision domain into two.
Switch: An intelligent Layer 2 device that uses an internal MAC Address Table to inspect incoming frames and forward them exclusively to the specific physical port connected to the destination device. Every individual port on a switch is its own collision domain.
Router: A Layer 3 device that connects completely separate networks. It reads logical addresses (IP Addresses) and utilizes internal routing tables to determine the optimal path for forwarding data packets across different networks.
Twisted-Pair Copper Cabling: Conductors twisted together to cancel out electromagnetic interference (EMI) from external sources.
Unshielded Twisted Pair (UTP): Most common for local area networks (LANs). Categorized by performance (e.g., Cat5e, Cat6, Cat6a, Cat7, Cat8), supporting speeds up to 40 Gbps over limited distances.
Shielded Twisted Pair (STP): Features metal foil wrapping around conductors to provide extra protection against high EMI environments.
Coaxial Cable: Features a central copper conductor surrounded by an insulating layer, a braided metallic shield, and an outer jacket. Historically used in bus networks; now primarily used for cable internet and television (DOCSIS standards).
Fiber-Optic Cabling: Transmits data as pulses of light through a core composed of pure glass or plastic. Immune to electromagnetic interference and offers massive bandwidth capacity over long distances.
Single-Mode Fiber (SMF): Narrow core (approx. 9 microns) that carries a single ray of laser light. Engineered for long-haul telecommunications stretching over kilometers.
Multi-Mode Fiber (MMF): Wider core (approx. 50–62.5 microns) allowing multiple modes or paths of light emitted from LEDs. Best suited for short distances within data centers and enterprise buildings.
Wireless Media: Employs radio frequencies (RF) or infrared waves to transmit data through the air without physical tethering. Utilizes standards such as IEEE 802.11 (Wi-Fi), cellular networks (LTE, 5G), and Bluetooth.
Network Designing is the engineering process of planning, sizing, and structuring a network deployment to meet specific operational requirements, performance parameters, and scalability objectives.
To manage complexity, large enterprise networks are structured using a highly scalable, modular three-layer approach:
+-----------------------------------------------------------------------------+
|Core Layer| <- High-speed backbone switching
+-----------------------------------------------------------------------------+
| |
+-----------------------------------------------------------------------------+
|Distribution Layer| <- Routing, policies, ACLs, VLAN routing
+------------------------------------------------------------------------------+
| |
+------------------------------------------------------------------------------+
|Access Layer| <- End-device connectivity (PCs, printers)
+------------------------------------------------------------------------------+
1. Core Layer: The backbone of the network. Optimized strictly for high-speed, low-latency switching of massive volumes of traffic. It must not perform any packet filtering, access control list (ACL) evaluations, or complex routing tasks that could slow down packet delivery.
2. Distribution Layer: The boundary layer between the access and core layers. It handles smart routing, security policies, packet filtering via Access Control Lists (ACLs), Quality of Service (QoS), definitions of Virtual Local Area Networks (VLANs), and inter-VLAN routing operations.
3. Access Layer: The edge of the network where end-user devices (workstations, servers, IP phones, access points) connect directly to the infrastructure. Focuses on device connectivity, port security, and network segment isolation.
IP Subnetting: The process of logically dividing a single large IP network block into smaller, distinct sub-networks (subnets). This controls broadcast traffic, optimizes address utilization, and simplifies security management.
Classless Inter-Domain Routing (CIDR): Replaced the legacy classful addressing system (Class A, B, C). It allows flexible allocation of IP addresses using a variable length bitmask notation (e.g., /24), matching network boundaries precisely to organizational needs.
Virtual Local Area Networks (VLANs): Logical network segments created within a Layer 2 switch environment. VLANs isolate broadcast domains without being restricted by the physical location of devices. A router or Layer 3 switch is required to route traffic between different VLANs.
Designing reliable networks requires removing single points of failure so that the system remains online even if specific hardware or lines fail.
Link Aggregation (EtherChannel / LACP): Combines multiple physical Ethernet links between two devices into a single logical channel. This provides both combined throughput scaling and automatic failover protection if a physical cable fails.
Spanning Tree Protocol (STP / RSTP - IEEE 802.1D/w): A critical Layer 2 protocol designed to prevent catastrophic loops in networks with redundant paths. It dynamically blocks specific redundant ports to ensure there is only one active logical path between any two nodes. If an active path fails, it recalculates and unblocks the redundant port.
First Hop Redundancy Protocols (FHRP): Protocols like HSRP (Hot Standby Router Protocol) or VRRP (Virtual Router Redundancy Protocol) that configure multiple physical routers to act as a single virtual "Default Gateway" for end-host configurations, ensuring seamless failover if the primary router goes offline.