- Computer Networking Course
- Computer Networking Tutorial
- Network Interface Unit
- Network Switch and Types
- Network Transmission Media
- Twisted Pair Cable
- Wireless Transmission Medium
- Satellite Microwave
- Data Communication Terms
- Types of Networks
- Network Topology
- Network Devices
- LAN Design
- Network Components Checklists
- Communication Protocols
- Mobile Computing
- Inter-networking Terms
- URL and Domain Names
- Web Hosting
- Web Scripting
- Software Categories
- Network Security
- Computer Programming
- Learn Python
- Python Keywords
- Python Built-in Functions
- Python Examples
- Learn C++
- C++ Examples
- Learn C
- C Examples
- Learn Java
- Java Examples
- Learn C#
- Learn Objective-C
- Web Development
- Learn HTML
- Learn CSS
- Learn JavaScript
- JavaScript Examples
- Learn SQL
- Learn PHP
LAN Design
When we discuss the performance of a network, the design of the network is the factor that has the greatest impact on that performance.
The following are some of the steps that will walk you through the process of designing the networks in general. In general, the problem of network design can be broken down into the three components that are outlined in the table below:
| Element | Description |
|---|---|
| Environment given | The environment provides a description of the physical locations of hosts, servers, terminals, and other end nodes, as well as the anticipated volume of network traffic and the associated costs for delivering the various tiers of service. |
| Performance constraints | Reliability of the network, traffic throughput (which is a measure of performance), and the speeds of both the host computer and the client computers are examples of performance constraints. Some examples of performance constraints include network interface cards and the speeds at which hard drive access is performed. |
| Internetworking variables | The network topology, line capacities, and packet flow assignments are all examples of variables that can be found in internetworking. |
The objective is to deliver a service that does not compromise capacities or packet flow assignments while at the same time reducing costs as a result of these factors as much as possible.
Good Network Design - The 80/20 Rule
The placement of clients in relation to the servers is critically important to the design of a successful network.
Client computers should, if at all possible, be placed on the same logical network as the servers to which they have the most frequent access. A logical connection can be defined in your network software so that users in one area of a building can be in the same logical network segments as a server that is located at the opposite end of that building. This is in contrast to a physical network connection, which would mean that a client and server were attached to the same hub. A physical network connection is the most common type of network connection.
This straightforward activity lightens the load on the network backbone, which is responsible for transporting traffic between the various segments.
In a properly designed small-to medium-sized network environment, not more than 20 percent of the network traffic should need to move across a backbone (the spine that connects various segments or subnetworks), while 80 percent of the traffic on a given network segment is local (destined for a target in the same workgroup), and 80 percent of the traffic on a given network segment is local (destined for a target in the same workgroup).
Congestion on the backbone may be an indication that the current traffic patterns do not adhere to the 80/20 rule. In this situation, it is simpler to improve network performance by implementing one of the following methods than it is to add switches or upgrade hubs. These methods are as follows:
- Transferring resources from one server to another, including applications, software programs, and files.
- Transferring users locally, if not physically, so that the workgroups more closely reflect the actual traffic patterns.
- Including the servers so that users can access them locally without having to travel across the backbone.
After you have ensured that you have a proper network design and that resources are located in the appropriate places, it is now time to determine the best technology to use in order to meet your growing needs.
Making Use Of Repeaters
As you are already aware, the fundamental unit that is used in networks to connect individual segments is called a repeater.
The 5-4-3 rule is a fundamental guideline for making use of repeaters. There shouldn't be more than three populated segments on the path that connects the two stations on the network; the maximum number of segments between those segments should be five, and there should be four repeaters in between those segments.
Making Use Of Bridges
Bridges, as you are already aware, are utilized in the process of connecting separate parts of a network.
Bridges are not the same as repeaters because bridges are intelligent devices that operate in the data link layer of the OSI model. Repeaters work in the physical layer of the OSI model. The collision domains on the network are under the control of the bridges.
Bridges are able to discover not only the MAC layer addresses of each node on each segment but also the interface on which that node is located.
The only time bridges will forward an incoming frame is if the MAC address of the destination device is located on another port or if the bridge is unaware of the device's current location. This type of situation is known as flooding.
Bridges filter any incoming frames with destination MAC addresses that are on the same segment as where the frame arrives. If they find such a frame, they discard it rather than passing it on to the next segment.
The spanning-tree protocol is used to create a network topology that is free of loops, and bridges are responsible for implementing this protocol. Because of this, it is possible that one or more bridges on a network will be blocked if the bridges are forming a loop.
Bridges are able to talk to one another and share information with one another by exchanging things like priority and bridge interface MAC addresses. They begin by deciding which bridge will serve as the root, and then proceed to implement the spanning-tree protocol.
On some bridges, particular interfaces are set to the hold state, while on others, those interfaces are set to the forwarding mode.
In accordance with the spanning-tree protocol, I would like to bring to your attention the fact that the network does not currently contain any loops.
When Is the Appropriate Time to Use Hubs, Bridges, Switches, and Routers?
It is not difficult to set up a network when there are only two computers involved. You will only need to install network interface controllers, or NICs, on both computers before connecting them together with a crossover cable.
And you should use the repeaters if you need to increase the distance that separates the computers in any way. Signals will be amplified and retimed by the repeaters on their own. As long as a company only required two computers to be networked, repeaters were perfectly acceptable.
And in the event that a company wanted to connect a third computer to the network, or a fourth computer, or any number of computers beyond that, then all they needed to do was use a multi-port repeater known as a "hub."
Concerns with the hubs
Hubs distribute available bandwidth among all of the devices that are attached to them. The traffic cannot be filtered by the hubs. They simply sent the data to all of the devices via broadcast.
When data frames are broadcast, they are sent to every connected node in the network. The utilization of a more intelligent hub known as Bridge is the answer to this predicament. Bridges have the ability to filter traffic on a network based on MAC addresses.
Now that the bridges are in place, traffic will only be directed to the appropriate segments rather than all of the segments as was the case with the hubs. The spanning-tree protocol prevents bridges in a network from connecting in a circular fashion so that a loop cannot be created.
It is possible to effectively replace the four bridges with a switch (or a multi-port bridge). Not only does it take the place of the four bridges, but it also assigns a dedicated bandwidth to each individual LAN segment.
Concerns with the Switch
A switch is unable to prevent data traffic (packets of information) from traveling to all of the other LAN segments even if they are destined for a different LAN segment.
A router is the answer to this problem's predicament. Routers filter the network traffic based on the IP addresses of the devices connected to the network.
The router can determine which local area network (LAN) segment the network traffic or data packets belong to based on the IP address.
Combining Hubs, Switches, and Routers
Hubs, which come in both Ethernet and Fast Ethernet varieties, find their most common application in relatively small networks, where there are only a few nodes on each segment. Hubs do not have the ability to filter collision domains on the network, nor do they control the broadcasts that are sent out. If you need a higher bandwidth, then you should use hubs that are 100 Mbps.
Move on to the switches once the number of nodes on the network has grown to a certain point.
Since the cost of switch ports is comparable to that of hubs, it is recommended that switches be used as the primary devices for network connectivity on the network. By establishing a collision domain for each port, switches eliminate collisions and alleviate the strain caused by media contention on a network. Switches should be used instead of hubs in environments where utilization is greater than 40 percent on Ethernet networks and greater than 70 percent on Token Ring and FDDI networks.
Routers should be used instead of switches when troubleshooting issues related to protocols because switches are unable to resolve broadcast characteristics of protocols.
« Previous Topic Next Topic »