Networks and Communication

Networking devices

• Switch: A device that connects multiple devices on a network. Generally operates using MAC (Media Access Control) addresses, which are specific to hardware.
• Router: A device that connects networks by forwarding data packets between them. Generally operates using IP (Internet Protocol) addresses, which are more generally specific to networks than a MAC address.
• Firewall: A firewall is a system which prevents unauthorised access of a system from the outside. For instance, it could prevent packets from entering through an unused port.
• Modem: Stands for Modulator / Demodulator. A modem is a device that encodes/decodes digital information for transmission / to decode the transmitted information.
• Network Interface Card: Also known as Network Interface Controller. A hardware component that connects a computer to a network. Many modern computers will have an NIC for Ethernet, an NIC for wireless and an NIC for Bluetooth.
• Wireless Access Point: A hardware device allowing a Wi-Fi enabled device to interface with a wired network, by transmitting signals wirelessly. A Wireless Access Point may be integrated into a router.
• Bridge: A bridge connects subnetworks to form a larger network. The difference between a bridge and a router is that a router simply allows multiple independent networks to communicate whereas a bridge creates a larger network containing both of the old networks.
• Gateway: A device which is capable of joining two networks which use different protocols. The gateway thus functions as an entry point into a network from the Internet, as well as being an integral component of most routers. Typically a firewall will be integrated into a gateway as the network's entry point.
• Repeater: A device that receives a signal and retransmits the same signal. This can be used to extend signals and thus allow it to be transmitted over large distances.

Transmission Media

Transmission media are physical processes used to transmit data from one device to another in a network.

Some examples of wired transmission media are radio, satellite, microwaves, and cellular data.

Some examples of wired transmission media include the following:

• Twisted Pair Cables: These contain thin copper wires arranged into twisted pairs; the pairs are twisted to minimise interference. An Unshielded Twisted Pair (UTP) contains two insulated cables together in a single cable, whereas in a Shielded Twisted Pair (STP) the pairs each have foil shields around each of the pairs and then there is an overall shield around all of the pairs in the wire. (There may of course be multiple pairs in a cable.)
• Fibre Optic Cables: These cables are made of thin strands of glass / plastic / a similar material that transmits information using light. A Fibre Optic Cable includes three main parts: the core, which is the thin glass centre where the light actually travels; the cladding, which is the outer material which reflects light back into the core, allowing it to be transmitted; and the buffer/coating, which is a plastic coating used to protect the internal fibre. A transmitter produces and encodes light; a receiver receives and decodes light. There are two main types of Fibre Optic Cables: single and multi mode. In a single mode Fibre Optic Cable, the core is approximately 9 microns (0.009 millimeters) in diameter. Single mode Fibre Optic Cables work over long distances and have a greater bandwidth. A multi mode Fibre Optic Cable, in comparison, has a core of about 50 to 100 microns, and works over shorter distances and have a high bandwidth (but not as great as a single mode Fibre Optic Cable). The main reason not to always prefer single-mode over multi-mode Fibre Optic is that single-mode is more expensive, normally about four times as much as multi-mode channels, due to an increase in the cost of the laser diode.

TCP/IP Model

The TCP/IP model was developed by the U.S. Department of Defence, and splits networking into four layers determining how messages should be transmitted across networks.

One advantage of the TCP/IP model is that people using different systems can communicate even if a different TCP/IP stack is used. In addition, layers may operate independently, so it does not matter strictly what technology is used at each layer as long as the protocol used for passing information is the same.

Layers of TCP/IP

The four layers are:

• Application layer: This defines the protocols for applications to communicate, and provides them with the ability to access other layers. Protocols on the Application layer include HTTP, SMTP and FTP. On this layer, only the data being transmitted is stored.
• Transport layer: This controls how links are established and is responsible for ensuring that packets have been sent across a network correctly and sequences them. Protocols used on this layer include TCP (Transmission Control Protocol) and UDP (User Datagram Protocol). This layer operates using port addresses. On this layer, information used by TCP is stored along with the original data. Firewalls operate on this level, since they are mainly concerned with ports.
• Internet layer: This controls how packets are addressed and routed across the network. Protocols on this layer include IP (Internet Protocol), and devices used on this level include the router. Devices are addressed by IP address, and extra information for use by IP is stored in this layer along with the information used by TCP and the original data.
• Network layer: This is responsible for placing packets on the network medium (e.g. cable) and receiving them off the medium, allowing the upper layers of the TCP/IP model to work with multiple different physical technologies. In this stage, information for use by the Network layer is stored along with all that which is used by the Internet layer. Protocols used include Ethernet 802.3 and 802.11x, and devices include the switch, Network Interface Card, and Wireless Access Point. Devices are addressed by MAC address at this stage.

Communication Standards and Protocols

To communicate, devices must have a common protocol for interpreting data that is being transferred.

Protocols for wireless communication include:

• Ethernet 802.11x: A set of standards for wireless local area networks allowing transmission of data wirelessly.
• Bluetooth: A wireless standard for technology allowing data exchange over short distances, generally used for low-power consumption using a short range based on relatively cheap transceiver microchips in each device. Devices do not need a visual line of sight since Bluetooth uses a radio broadcast system.
• RFID: Radio-frequency identification. Uses electromagnetic fields to transmit / detect information.
• WAP: Wireless Application Protocol. A standard for accessing information over a network.

Protocols for wired communication include:

• Ethernet 802.3: A collection of standards defining media access control in wired Ethernet.

IP addresses may be assigned via IPv4 or IPv6. The difference is that IPv4 uses 4 bytes (32 bits) per address but IPv6 uses 16 bytes (128 bits).

Network Control Protocols

A Network Control Protocol is designed to control traffic flow across a network, ensuring that only one device may transmit at a time.

The two most common network protocols are CSMA/CD and CSMA/CA.

CSMA stands for Carrier Sense Multiple Access. This means that each device can listen to what is occurring on the network prior to transmitting data, and that multiple devices can listen to network traffic at one time.

CD stands for Collision Detection, indicating that each device can detect when there is a collision and retransmit after waiting for a random period of time. CA stands for Collision Avoidance, so when a device is about to transmit, it sends a signal telling the other devices not to transmit, and waits long enough for other devices to receive the signal; if it receives this signal from another device of course it will not begin transmission.

Error Checking

There are two methods commonly used in network communications to detect errors.

• Parity bits involve using an extra bit at the end of data such that the number of 1s in the data including the parity bit should either be even (or odd, if that is how it is implemented).
• Checksums involve performing some algorithm upon the data which should approximately uniquely identify the data, such that there is a very low probability that the data is transmitted incorrectly and the checksum remains the same.

Data safety / security

An important concern with regards to a network is its security.

The following are common important security concerns:

• Denial of Service attacks: These involve preventing users from accessing a network resource by disrupting services of a host. This is typically done by flooding the target resource / machine with requests to overload it. Prevention methods may include the following: using a firewall to restrict bandwidth usage to only authenticated users, and investing in third-party services which protect against DoS attacks.
• Back doors: A back door is a method of bypassing encryption or authentication. Almost always these are unintentionally left in code / accidentally inserted into code, which poses the main security concern; alternatively they may be intentionally inserted into code as a method for an administrator to perform administrative actions. It is difficult to protect against a backdoor attack since their existence is typically not known to those maintaining the network, but it is possible to take some preventative measures. Firstly firewalls can be implemented to only allow entry by authorised users. Secondly, anti-malware solutions may be used to prevent programs from being surreptitiously inserted into a network. Thirdly, networks can be monitored to ensure that malware is not present on the network.
• IP spoofing: IP spoofing is the creation of Internet Protocol packets with a false source IP address, for the purpose of hiding the sender's identity, or possibly impersonating another user. IP spoofing may be used in conjunction with a DoS attack to make the attack less obvious (since it appears that it is coming from multiple users). Preventative measures include using authentication based on key exchange between machines on a network; using an access control list to determine which IP addresses may be used to access the system; implementation of filtering inbound and outbound traffic; configuration of routers and switches to reject packets outside of the local network; and enabling encryption session so that trusted hosts outside the network may securely communicate with local hosts.
• Phishing: Phishing is the attempt to obtain sensitive information by impersonating a trustworthy entity, based on the word fishing due to the similarity to using bait to catch fish. To prevent phishing one can raise user awareness of how to recognise phishing and take security measures to disallow phishing attempts.

Network Storage

There are two main types of network storage.

• NAS: Stands for Network Attached Storage. This involves the use of specialised storage devices attached to a network, providing file-based storage to other devices on the network. Typically these devices will use a small, embedded OS. NAS units connect using the regular TCP/IP model, and units are given IP addresses to connect with other devices.
• SAN: Stands for Storage Area Network. This is a network dedicated to transportation of data for storage and retrieval. Storage devices are attached directly to the network. SANs only carry I/O traffic between the servers and storage devices, not application traffic. This allows shared storage arrays to be interfaced with by multiple servers.

There are two common types of SAN.

• Fibre Channel: This is a set of standards for connecting devices in a 'fabric network'. Each server needs a specialised network card, which connects to a Fibre Channel switch, or a 'fabric'. The specialised devices mean that there is a greater expense to set up the system as well as maintain it. Fibre Channel SANs connect multiple servers to storage, and is optimised for the transfer of large blocks of data.
• iSCSI: This is a protocol based on TCP/IP for establishing and managing the connections between IP-based storage devices, and is able to use regular technologies such as NICs, switches and Ethernet technologies.

The following are some significant diifferences between NASs and SANs.