Structure and Design of a Network
Posted date: June 09, 2017
Paper Title: Network Structure and Design – When an organization wants to implement a network, it is necessary to consider a number of design factors before going ahead. First of all, the organization’s network requirements need to be defined in a very clear manner and every required network component must be identified. The following article looks at the key considerations for such a project.
Defining the Architecture and Design Features of a Network
The term “network architecture” is a reference to the infrastructure upon which a network is built. This usually comprises of specialist software, equipment for transmitting data, and specific communication-based protocols that define the logical and structural layout of the actual network itself. Wireless or wired transmission methods may be used in a network with much dependent on an organization’s specific requirements. A network can be any one of a number of different types, with the size of the network often being a deciding factor. A LAN (local area network) is a computer network that serves a small and localized area e.g. within one office or building. By contrast, a MAN (metropolitan area network) is a network that links computers within a given city, and a WAN (wide area network) is a network that connects computers over a wide geographical area, even across a whole country or between countries. Of these three network types, an organization is likely to implement a local area network since this covers a small, local area.
Media for Transmission
The medium by which data is transmitted across a network may be of the wired variety or the wireless variety. A wired network means using fiber-optic or coaxial cables while a wireless network is a method of transmission that does not involve any wiring. Requirements related to throughput, good put and bandwidth are the factors that generally decide the most suitable transmission method. Because fiber-optic cabling is not prone to collision, signal loss is low, and this method is efficient for transferring data across a network that handles high volumes of traffic. On the other hand, although coaxial cables cost less than the fiber-optic variety, these are prone to collisions so signal loss can be high. In LANs where there are not many computers, wireless networks are an efficient option.
Network Management Methods
There are two primary methods for managing a network, and these are a) client-server and b) peer-to-peer. The first method, client-server, is a layout where a centrally located server provides various services to a number of independent clients. In the peer to peer layout, communication occurs between a number of connected computers without the help of a central server. In the latter type network, several computers are able to share any one application that is installed on one of the member computers. The client-server arrangement is designed to serve and support several clients where the latter do not share any resources. In fact, security is better in a client-server network since the server handles security. Moreover, this model is easy to expand and upgrade in the event an organization has new or changed requirements.
Topology of a Network
It is possible to divide a network’s topology into the logical and physical types. The former – logical topology – refers to how the data transmitted via a network is laid out. The latter type - physical topology – is a reference to how computers and other network devices are inter-connected. The primary topology types are:
- Ring topology: This type involves the chosen cable type being run around the area where nodes are connected to one another. This topology suffers less collision than the bus variety.
- Bus topology: This type is a set-up where all the devices on a network are linked by just one cable. This topology is suitable for smaller networks, but collisions are frequent and speed (slowness) is an issue. Collision is reduced or avoided with the use of token rings.
- Mesh topology: This type of topology is an arrangement where all devices on a network are connected to one another for the purposes of redundancy, fault tolerance, and improved performance.
- Star topology: This type of topology is an arrangement where all devices on a network are connected to one centrally-installed hub. This centralized management method accommodates faster upgrading, but if the central hub fails, the whole network comes down.
Security Considerations When Designing a Network
Hackers and others with malicious intent often attack computer networks. Because of this, network security is a key consideration in the design stage. In order to reduce, or preferably avoid, the number and frequency of attacks, it is essential for a network to be built with:
- An intrusion detection system
- A firewall
These are the primary measures used to detect unwanted intruders in a network and to reduce and/or eliminate threats.
Network Should be Scalable
Scalability is a reference to being able to grow or extend a network. A network should be sufficiently scalable to be able to accommodate growth in an organization and its network/communications infrastructure.
NAT (Network Address Translation)
In order to enhance network security, Network Address Translation should be considered in the design stage where several computers access a private network using a single IP address.
Network Design and Open Systems Interconnection (OSI) Reference Model
A seven-layer reference model or framework, as described below, OSI is concerned with how applications communicate with each other across a network. In this case, the system is divided into layers - whereby the top layer sends and receives services requests to and from the next layer down. Each layer in turn performs the same function in respect of sending and receiving service requests to the layer above and beneath it.
The seven layers are:
- Layer Name: Physical - This first (physical) layer is a reference to the component hardware and all the devices connected to a given network. This layer is used to define all physical components (devices) and transmission methods. It is also the recipient of service requests from the second layer down (the data-link layer). As another function, it encodes and decodes signal data. This layer includes Ethernet and CSMA/CD protocols.
- Layer Name: - Data-Link - This second (data-link) layer receives and sends service requests from the layer above (the physical layer) and to the layer below (the network layer). The primary function of this layer is to deliver data in a reliable manner over a network. It additionally performs such functions as detecting and correcting errors, controlling error and flow, and framing. The data-link layer has two sub-layers, which are a) logical link control and b) media access control. The former – logical link control – checks for errors, controls flow, and manages package synchronization. This layer includes ATM and X 25 frame relay protocols.
- Layer Name: Network - This third (network) layer has responsibility for the management of all a network’s connections, packet routing from source to destination, and congestion. The network layer gets services requests from the layer below (the transport layer) and sends similar type requests to the layer above (the data-link layer). This layer includes ICMP, IGMP, and IP protocols.
- Layer Name: Transport - This fourth (transport) layer is responsible for delivering error-free data reliably across a network. It achieves this by detecting and correcting errors. This layer prevents data loss, and it receives data in the usual manner i.e. from the layer below. The transport layer is also responsible for providing connection-oriented or connectionless service.
- Layer Name: Session – This fifth (session) layer serves the primary purpose of establishing and terminating network sessions. Essentially, it sets up and ends connections between two processes or more. It is additionally responsible for managing inter-host connections. The session layer is responsible for the network validation step where login and password validation is required, and it also provides a check-point method i.e. where errors occur, data is retransmitted from the previous check-point. This layer includes RIP, SAP, and SOCKS protocols.
- Layer Name: Presentation – This sixth (presentation) layer has responsibility for compressing and decompressing data, manipulating data, and presenting data. This layer is the recipient of service requests from the layer below (the application layer) and sends requests of a similar nature to the layer above (the session layer).The presentation layer is further involved in the semantics and syntax of in-transmission data and, for security purposes, it uses data encryption methods and decryption methods or cryptography. This layer includes ASCII, EBCDIC, JPEG, MIDI, and MPEG protocols.
- Layer Name: Application - This seventh (application) layer is responsible for end-user interactions and provides such services as email and file transfer services. The application layer transmits service requests to the layer above (the presentation layer). This layer’s communication methods include DNS, FTP, HTTP, NFS, SMTP, TELNET, and TFTP protocols.
Protocols Used in Network Design
- ARP (Address Resolution Protocol): This protocol provides physical addresses known as IP addresses.
- DHCP (Dynamic Host Configuration Protocol): A protocol used for controlling and managing the IP addresses in a particular network.
- DNS (Domain Name Service): This service provides domain names for given IP addresses.
- Ethernet: This accommodates the transfer of data/information via Ethernet cable from one physical location to another.
- FTP (File Transfer Protocol): A standardized method for the transfer of files from one host to another across a TCP-enabled network.
- HTTP (Hypertext Transfer Protocol): A standard-type method for transmitting documents via the Internet.
- ICMP (Internet Control Management Protocol): This is used for managing errors while data and packets are being transmitted between computers.
- IGMP (Internet Group Management Protocol): Provides multi-casting support.
- IP (Internet Protocol): This protocol facilitates packet assembly, packet fragmentation, and packet routing.
- PPP (Point-to-Point Protocol): This protocol is an improved version of SLIP, and its function is encapsulating serial line data.
- RIP (Routing Information Protocol): This protocol is used by routers for routing inter-network information.
- SLIP (Serial Line Internet Protocol): Used for encapsulating serial line data.
- SNMP (Simple Network Management Protocol): A protocol for managing all the elements of a network based on received and sent data.
- STMP (Simple Mail Transfer Protocol): A standard-type method for transmitting mail messages from one server to another.
- TCP (Transport Control Protocol): This protocol delivers packets in a reliable and connection-oriented manner.
- Telnet: Used to access remote (computer) devices.
- UDP (User Datagram Protocol): This protocol also delivers packets but in a non-reliable and connection-less way.
- First Layer - Network Access: This first layer places TCP/IP packets into the network medium and receives packets from the same channel. It additionally controls all hardware and devices in a network, and it combines both the OSI model’s physical (first) layer and the model’s data-link (second) layer.
- Second Layer – Internet: The Internet layer functions as the OSI model’s network (third) layer. Its function is to route and to provide a network’s addressing and also its packet addressing.
- Third Layer – Transport: The transport layer here performs the same function as its OSI counterpart layer. The primary function is delivering error-free data reliably. The transport layer receives and sends service requests from the application (fourth) layer and to the Internet (second) layer.
- Fourth Layer – Application: This (fourth) layer contains applications for interacting with end users. The layer’s composition is a combination of the OSI model’s fifth, sixth and seventh layers i.e. the session, presentation and application layers respectively.
Troubleshooting Network Issues - The TCP/IP Commands Commonly Used
The following are some of the TCP/IP commands commonly used to highlight glitches in network communication:
- ARP: Used for viewing and editing address resolution protocol cache.
- HOSTNAME: Used for showing and displaying a computer’s host name.
- PING: Used to test network availability or reachability by sending ICMP echo.
The event viewer displays a list of all events and errors.