Network Hardware

In addition to all the instructions that protocols provide, there are dedicated pieces of hardware that make a network work effectively and efficiently. These pieces of hardware are all advancements of the ethernet network protocol. Most network hardware today is based on the twisted pair or fiber cable connection medium.

Hub

A hub serves as a central node or device in a computer network. A hub works to connect multiple ethernet segments, computers, or other devices together and treats them as a single segment. A hub provided the first method for which computers could share a single connection to the internet. Hubs work at a maximum speed of 100 million bits per second or 100Mbps.

A hub is used to connect multiple computers together. It connects computers together through what is known as a shared medium. A good example of a shared medium you may have encountered is a one lane bridge. Though many cars can use a one lane bridge, and cars can travel across the bridge in either direction, only one car can travel across the bridge in a single direction at any given time. In much the same way, only one computer can talk to the hub at a time, and that computer cannot both send and receive data at the same time. Now, a computer can send or receive data really quickly such that it seems like multiple computers are sending and receiving at the same time, but in actuality, they are not.

One disadvantge of a hub is its unintelligent design. A hub cannot distinguish who or what is connected to each of its ports. Thus, when it receives data from a given connection (see Fig. 1), that data is broadcast to each connected segment/device (see Fig. 2). This method of communication is what allows the hub to resemble a single segment. It is up to the device to either accept or reject the pieces of data. This constant broadcast to all parties not only creates collisions of data that can interrupt and degrade service, but also poses a security risk.

Fig. 1: Packet of data sent to hub

 

Fig. 2: Packet of data broadcast to all connected nodes.

Switch

A switch is similar to a hub in that it connects network segments together. However, switches do not use a shared medium to connect their segments together. They separate all of their connections so that if you have computers A, B, C, and D connected to a switch, the switch will allow computers A and B to send and received data to and from one another at the same time as allowing computers C and D to send and received data to and from one another. As you can imagine this allow data to travel between computers much faster that using a hub can function with its shared medium.

Switches also track which computers are connected to which of its segments. This allow it to send traffic to only an intended recipient. So, if computer A sends the switch a packet of data intended for computer C, the switch will only deliver the packet to computer C, unlike a hub that will send the packet to all computers connected to it. This is referred to as intelligence, meaning a switch is a more intelligent device than a hub. This provides for a more secure network because computers B and D cannot listen in on what computer A sent to computer C. Because switches are faster and more secure, they generally replace hubs in most networks.

Fig. 3: Packet sent to switch

 

Fig. 4: Packet sent to addressed node

 

The downside of using a switch is space requirements.  Instead of relying on devices to interconnect themselves, switches are used to connect the various devices.  The more complex the network, the more switches are needed to connect the various devices together.

Router

A router is a specialized computer that controls how data moves between networks, not within networks.  Routers ensure that data goes where it is supposed to go and not where it is not intended. While switches deliver information that they are given to specific paths, routers can determine what path that data should follow based on a routing configuration table. This table keeps track of all connections to the router and all possible paths that exist between networks. This allows the router to choose the quickest/best path for data.

Different Kinds of Routers

• Internet Connection Sharing is when you have several computers connected together though a hub or a switch, but one of those computers is connected to the Internet in addition to the hub or switch. That computer can share its Internet connection with the other computers, and then acts as a router for the rest of the computers. In this case, when one of the other computers connected to the hub or switch needs to talk to a computer somewhere on the Internet, it talks to the Internet Connection Sharing computer which then routes the traffic on to the Internet. This is usually such a simple task that it is able to run in the background of one of the computers without affecting the processing power.
• Home broadband routers are used to connect multiple computers together in a home setting. They are individual devices rather than a normal computer and have specialized software just for routing data between the home network of computers and the Internet. Frequently, these devices will have other features built into them as well such as hardware that allows computers to connect to them wirelessly, software to perform NAT (discussed below), or software that acts as a firewall.
• Office Network routers are used to connect more than two computers together in an office setting. They are usually individual devices rather than a computer or server due to the high traffic expected with so many devices. These routers also contain some rudimentary security protocols to protect the network from low-level security attacks.
• Major Traffic Routers are used in nexus points on the internet that must examine and transfer huge amounts of data. Since extremely large amounts of information flow through these routers every second, these routers must be very powerful systems

With the proliferation of the internet over the last few decades, we have come across a problem -- we are running out of addresses for computers to use! The current numbers estimate that there are around 350 million users actively using the internet. There are projects going on now to increase the number of possible addresses for use, but such a project will take time to implement correctly. Until then, we have turned to Network Address Translation (NAT) to help solve the problem. A NAT router acts as a single address for incoming and outgoing data for a network, and will assign sub-addresses to each individual computer in the network.

For example, let us look at post office mail that comes to a CSC Faculty member. All mail, while addressed to an individual, comes to Campus Box 8206. The mail is then sorted internally and either placed in mailboxes for pickup by individuals, or hand delivered. A NAT router works the same way. All information for the internal network (usually a LAN) comes to the one address of the router, (see Fig. 5) and is then sorted internally with little outside knowledge (see Fig. 6).

Fig. 5: Data packet being sent to NAT router

 

Fig. 6: NAT router translating/forwarding data packet

 

There are two types of NAT addressing. One is static NAT. With a static NAT, a computer within the network always has the same IP address - for example, 192.168.1.115. This allows an outside device to access this computer more easily and readily. The other addressing schema allows for changing addresses. Dynamic NAT, which is most commonly used, assigns a computer an address when it connects to the network from a pool of available addresses. For example, a computer may have any address from the available IP address pool of 192.168.1.100 - 192.168.1.150. Today a computer could have the IP address 192.168.1.122 and tomorrow it could have 192.168.1.147.

Firewall

A firewall is simply a piece of hardware or software that works to prevent unauthorized communications from entering or exiting the network. The firewall will use a set of policies, or rules to govern whether or not a specific piece of communication is allowed to pass. Firewalls exist in two main forms. One is the personal firewall that is typically a piece of software that is run on an individual computer. Zone Alarm, Norton Internet Security, and Windows Firewall are all examples of such software. A more traditional firewall is an actual physical device that exists in between two networks. This device may be a computer or other intelligent device working to prevent unauthorized communication as it filters all traffic for the network.

Fig. 7: Data packets sent to node with firewall

 

Fig. 8: Firewall allows or blocks packets based on rules

 

Firewalls work at one of two levels. Many firewalls work at the network level where they are monitoring TCP/IP protocols. These firewall doesn't allow packets to pass if they don't meet the set of rules established by the firewall. Other firewalls work at the application level. These control whether or not specific applications (ie, a web browser such as Firefox) running on a computer can send or recive packets.