Network Design

Category: Technology


In the age of modern technologies, every company, be it a small company or a transnational corporation, is striving to use technologies to satisfy the needs of their customers. This paper is devoted to the case study of a small developing building company, the owner of which is willing to build up a network. The company is a new one, and the owner does not have the proper equipment for the network; that is why one of the key tasks is to make up a list of the devices that will be used. Therefore, the tasks of the developer are the following:

-   To create a local network that includes two laptops, two workstations, and one server that supports two users;

-   Besides, the owner wants the local network to have the access to the Internet, wireless connection, file share option, the option of creating multiple email addresses and the connection to the printer;

-   The customer should get a detailed plan of the network;

-  The customer is also willing to have a price list of the hardware and software (data security systems) that will be used.

Network Design

First, the developer should realize what kind of network he will use. Besides, the peculiarity of the future network is that the employees will use both Internet and Ethernet. Ethernet will be used for the in-company messaging system and inner websites, whereas Internet will be used for communicating with customers from all over the world. Ethernet networks are designed so that the packets are sufficiently large to fill the entire cable network signals up to the last bit of the transmitted information. That is why the Ethernet packets must be at least 64 bytes in length, the system before transmission complements the short bursts of up to 64 bytes, and the Ethernet standards imposed such strict limits on the length of the cable segments. The computer can detect a collision while it is in the process of data transfer. If you are using UTP cable or fiber optic cable, computer assumes that the collision occurred when the signal is recorded as a transmitter and receiving wires. In the coaxial networks of conflict it says the sharp surge in the cable. If the cable is too long or too short packet, the system can complete the transfer before the collision detecting signs. Collision is detected, the system immediately stops transmission and starts transmitting the signal jam (jam pattern), indicating all systems on the network that a collision has occurred; therefore, it is necessary to ignore all packets partially accepted by this time, and transmitting data was not begun until the network was cleared. After the jam signal the system pauses before trying to transmit data. The length of this pause is called the back-off period, and both systems involved in the collision calculate it by using the algorithm of the truncated binary exponential back-off, using a random number generator. Randomly period is calculated for two systems; it does not match, and they start transmitting simultaneously, creating new collision. Procedure steps CSMA / CD are such that the more systems you connect to the network and the more data they transmit, the more collisions occur. Collisions are a normal part of the Ethernet, but that does not mean that they do not interfere with it. Collisions cause delays in data transmission as some systems have to send the packets again. If the number of collisions is not too high, the delays are not significant, but when the network load increases, the number of collisions also increases, and the associated total delay significantly degrades the performance of the network. Therefore, the Ethernet is not recommended to get too overloaded. To reduce the load on the network, it is possible to install a switch or a bridge, or to split it into two LANs connected by a router. It may seem that control access to the network environment by using the mechanism of CSMA / CD is not effective; however, the actions by which the system is fighting for access to the network and eliminating the effects of collisions are so fast that the delay can be ignored if the number of collisions is not over the top. One of the main functions of the protocol in Ethernet is to encapsulate the data frame received from the network layer protocol and to prepare it for transmission over the network. A frame consists of header and trailer that are divided into fields that contain the information needed to deliver a packet to its destination.

In local networks that use the shared communication medium (for example, the local bus network topology and a physical star) it is significant that workstations have an access to the medium, as if the two PCs are beginning to transmit data at the same time, the network is a collision. In order to avoid these collisions, a special mechanism that can solve this problem is required. The bus arbitration is a mechanism designed to solve the problem of collisions. It establishes the rules that define the workstations when the medium is free, and can transmit the data. There are two methods for bus arbitration in Ethernet:

  • collision detection;
  • token passing.

Collision Detection

When the method of collision detection is used in local networks, a computer first senses and then transmits. If the computer senses that the transfer is made by someone else, it must wait until the data is transmitted and then take the second attempt. In this situation (two computers transmit at the same time) collision detection system requires that the sending computer senses the channel and continues to look for information that is sent by other people; if it is found, the transition should be stopped and resumed after a while. Sensing the transmission channel is called carrier sense, and sensing during transmission is called collision detection. PCs coming this way use a technique called "collision detection with sensing carrier", abbreviated CSCD (Virtualization Design Guide).

Token Passing Local Networks

Token-passing systems work differently. To transmit data, a computer must first obtain permission. This means that it must "catch" the data packet of a special type called a token that is circulating in the network. The marker moves in circles, passing successively each network computer. Each time the computer has to send a message, it catches and holds the token at home. Once the transfer is finished, it sends a new marker on a journey down the network. This approach ensures that any computer sooner or later will have the right to catch and hold the token as long as its own transfer is completed.

The point is that developers of the network had to decide what kind of network should be used. There are four main types of configurations of workstations in the network. These species are referred to as topologies.

- "Bus". These networks consist of a common cable (or bus duct) to which all clients are connected. To prevent signal leakage or reflections at the ends the so-called line terminators (absorbers’ signal) are arranged. This system uses a relatively simple mechanism of data transfer: sending a computer message applies to all members of the network, each of them in turn determines the message being addressed to them or not. It uses a kind of coded signal (signature). “Buses” by definition involve the same rights and devices of all workstations, but because there is no server, there are no other users with different rights. The advantages of this scheme are low cost and easy installation, the network cables, and an easy set up. The ease of use is caused by the fact that the failure or shutdown of any workstation does not affect the others. The disadvantages include the need to reduce the performance with the addition of new workstations as well as the vulnerability and complex troubleshooting. These networks are vulnerable because the simple cable break or failure of one terminator completely destroys the system. This topology is used in local area networks with the architecture of Ethernet (10Base-5 classes and 10Base-2 for the thick and thin coaxial cable, respectively). The advantages of bus topology networks:

  1. failure of one of the nodes does not affect the whole network;
  2. network is easy to set up and configure;
  3. network is resistant to faults of individual nodes.

       The disadvantages of bus topology networks:

  1. cable break can affect the entire network;
  2. limited cable length and the number of workstations;
  3. it is difficult to identify defects in the compounds.

- "The Ring." A computer enclosed in network technologies can communicate with only two neighboring computers: one receives and transmits more, because in such a network need not terminated. Usually, all the computers in the network have the same rights, but sometimes it can be a special user who manages or controls the exchange. Of course, the existence of such a special party reduces the system efficiency since its immediate failure weakens the whole system. Among the advantages of this technology we must note the simplicity of installation and lack of need for additional equipment. Also, such networks provide stable operations without falling of the speed of data transmission and reception even when the network is fully loaded. Of course, computer networks also have their pitfalls. These include the complexity of configuring each computer for the smooth operation and the complexity of troubleshooting and fault as well. As it has already been mentioned, failure of one station is deteriorating work of the whole system.

- The most common computer network topology today is a star. The essence of this topology is that the workstations are connected to one central node, which is often a switch. Such a network may be either separate or a segment of a larger network. All communication is provided through the main computer (server) that is specially adapted for this purpose. The star topology became widespread because of the following qualities:

  1. Easy scalability (the network can be easily expanded);
  2. Easy troubleshooting;
  3. High performance;
  4. Good administration features.

However, this topology has such drawbacks as the need for a large amount of cable. Moreover, the number of possible workstations limits the number of ports on the main hub. If it fails, the entire network will be unworkable.

- The last of the existing topology is mesh topology. Its main peculiarity is the full-mesh, that is, every computer of the network is connected to all the others. This topology is characteristic mainly of large networks, since it can combine a lot of individual workstations. Often, these networks make wireless as well as cable networks to connect all the computers in the first place. It is very difficult and very expensive. The main advantage of this type of networks is the reliability of the connection, since in case the cable breaks with a particular workstation at a computer, there are still many ways of communication. (Networking Infrastructure and Design).

In this case, the developer will use the star type of the topology of the local network; however, the topology will be a bit modified, because the customer wants to have both Ethernet and Internet connection. In this case, the developer cannot use the Token ring technology because it is rather expensive and the budget of the customer is limited.

The following devices will be used:

Wireless technology will be performed by the Cisco Aironet 2600, the most advanced wireless access point. Its main features are the following:

  • It has two integrated antennas (2.4 GHz and 5GHz, both with the beam width of 360°);
  • It can be supported with all the LAN Cisco controllers;
  • It includes 3x4 MIMO with three spatial steams;

Cisco ASA 5510 SEC BUN K9 Firewall features:

  • Market-proven firewall, IPS, network anti-virus, and VPN capabilities

300Mbps Firewall Throughput, 170Mbps 3DES/AES VPN Throughput;

  • 250 VPN Peers;
  • 5 Interfaces;

Active/Stanby Failover, VPN Clustering/Load Balancing

Cisco 2900 Series Integrated Services Router:

  • The router provides such a high quality performance because its has the inbuilt Agile Application Services. Such services are able to host Cisco and multiple third party applications at the branch office; besides, it assures the high delivery performance, which support AVC (applications visibility and control) that can prioritize and optimize all types of data, voice, and video applications.
  • WAN high speed connection is achieved by the wide choice of protocols that may be used, such as T1/E1, T3/E3, 4G/LTE, xDSL, copper, fiver Gigabit Ethernet and others.
  • The optimization of WAN performance is made through the optimization of applications that are used through service modules.
  • Such technologies as VPN with IPsec, SSL VPN ensure the high level of data security and are enhanced by onboard encryption acceleration and threat defense support through firewall and intrusion prevention system (IPS) options, and further includes support for next-generation encryption and cloud-based security. This feature is significant for the company, because it is going to enlarge the amount of offices they have; besides, the company is planning to grow internationally, which means that every year the amount of private customer information will enlarge and it will need to be protected.
  • 2 RU and 1 RU units with up to 3 GE ports and up to 1 SFP port;
  • UCS-E service module that can support Cisco and third-party apps, VMware ESXi, and MS hypervisor;
  • Up to 2 service modules, 1 integrated service module (ISM) to host application services;
  • Up to 50 LAN switch ports, 4 Enhanced High-Speed WAN Interface Card (EHWIC) slots.

The company will use Dell laptops, as well as the workstations. The cost list will be:

Cisco Aironet 2600 - $1095

CISCO2901-SEC/K9 - $1459

Cisco ASA 5510 SEC BUN K9 - $1915

Dell Workstations (2) - $1348

Dell Laptops (2) - $1000

Besides, the company should include the prince of wires and further maintenance. It has been discussed that the price for maintenance will be $200 per month; however, the company will not manage to spend $5000 or even less for high quality equipment.

It is rather difficult to create and develop a network for a company with limited budget, that is why the developer can propose to buy only one workstation for the first period of the work of the company. It is also a rational proposition because the company will be located in one office at the beginning; therefore, it is reasonable to use only one workstation for now.

Related essays