Assignment 2 – Tobias Lotzke
Chapter 4
Question 1: Define local area network.
A local area network (LAN) is a small network with a small number of clients connected. It’s quite typical to separate LANs to constructional factors like one house or one building. The data transmission rate for LANs is about 10-100Mbps even though today’s technology allows speeds up to 1000Mbps.
Question 2: What does the data link layer do? What does the physical layer do?
The Data Link Layer ensures an error free link transmission. Using checksum-techniques like the Cyclic Redundancy Check (CRC), it provides a bit-level security for data transmission while only adding a really small overhead. This is one of the most important features in the OSI model because in most cases there is a high demand for error free systems.
The Physical Layer is responsible for sending data (in form of bits) over one wire. It uses different techniques (like Amplitude-, Frequency- and Phase modulation) of modifying a sound wave to deliver data to the next machine with the best effort.
Question 3: What is a NIC? What is a hub? What is a switch?
The Network Interface Cards (NIC) is the device which connects a computer to a network. Today, most machines already include NICs, but there is a possibility to upgrade a machine without a network card (using PCI for desktop systems or PCMCIA for mobile solutions).
A network hub has two main features: First, it acts as a bridge to connect different computers in a network. To join a network, the user just has to plug in the cable into one of the remaining ports. Second, the hub amplifies a signal. Due to the fact that the signals send through a network cable typically can only be transmitted for several meters, a hub is needed for larger distances. At any time the hub gets an input signal, it transmits this signal to all connected cables hence multiple transmission at one time cannot be processed.
A switch basically has the same features but differs in one point: It creates routing tables that prevent the switch from sending data to connected machines even though the data is not addressed to them. It reads the head of a frame and sends the data to the right receiver listed in the routing table. At the beginning, this table is filled by analyzing the incoming data and adding the source address.
Question 4: What is a topology?
A topology is the way how computers are connected within a network. There are different topologies like the star topology. They can be created by using a hub or a switch.
Question 9: How does fiber-optic cable differ from STP?
Fiber-optic cable is thinner and lighter. In addition, these cables ensure a higher speed. In contrast to that, they are way more expensive than and not as flexible as STPs because the glass can be damaged very easily.
Question 10: What are the different types of fiber-optic cable?
Multimode fiber (MMF), which means that the light is reflected within the cable and therefore increasing the time that a signal at the edge of fiber needs to arrive at the destination. As a result of that, they are limited to short distances. In order to reduce this effect, a new MMF, called graded index MMF, was created.
Single-mode fiber (SMF) transmits the data directly with only one stream of light. Therefore, no reflection occurs, which means that this signal can travel over long distances (up to 100 kilometers).
Question 11: In what ways is a hub similar to a switch? How do they differ?
Both create a network between the different connected machines and both amplify the incoming signals. But a switch routes the incoming data to the addressed machine whereas a hub sends the incoming data to every connected machine.
Question 14: How do layer-2 switches know where to send the frames they receive? Describe how switches gather and use this knowledge.
There are three methods of sending frames.
Store and forward: The switch buffers the data, performs a checksum and then sends the data to a machine in the network (using the routing table)
Cut through: Reads only up to the first address part (hardware address) and immediately sends the frame without checking for errors.
Fragment free: Combines both. Checks the whole address header in order to ensures that the packet will reach its destination. Error-checking is performed by the receiver.
Question 15: What distinguishes serial transmission from parallel transmission? Give an example of each on your computer.
The serial transmission sends a data stream byte per byte over one wire. An example is the USB (Universal Serial Bus) port, used by external hardware.
The parallel transmission sends the same data stream over multiple wires and separates the stream on to the different wires. An example is the old printer port (serial port) that was used to transmit the data to the printer.
Question 16: Briefly describe how CSMA/CD works.
Carrier Sense Multiple Access with Collision Detection is used to avoid collisions while two computers sending data at the same time. A computer waits for a small amount of time and listens to the network to make sure that no other machine is sending data. After ensuring that there is no transmission, the machine can start sending data. If two machines start sending at the same time, they will send a jamming signal to stop every computer from sending data. To ensure that the computers do not start at the same time again, they’ll wait for a random time to start transmitting the data.
Question 19: Explain the terms 10Base-2, 10Base-T, 100Base-T, 1000Base-T, 10GbE and 10/100 Ethernet.
10Base-T describes the earliest form of 10Mbps Ethernet. It uses twisted pair cables to transmit data with a maximum rate of 10 million bits per second (Mbps).It uses “Manchester encoding” to signalize when the machines can start or stop sending data.
10Base-2 basically uses the same system as 10Base-T but sends the data over copper wires which are way more expensive and therefore not used anymore.
100Base-T describes the faster way of transmitting data over CAT-5 twisted pair cables with 100Mbps. As a contrast to 10Base-T it sends data in packets of 5bits, 4 containing the data and the last one containing the signal that the packet is finished (called 4B5B coding).
1000Base-T runs over one cat 5 cable using parallel transmission. It supports speeds up to 1000Mbps and works with half-duplex.
10GbE supports speeds up to 10Gbps and uses optical-fiber cables. Due to the fact that it is full-duplex, it only runs in environments with circuit switched networks (only two computers connected at the same time). It uses parallel transmission.
10/100 Ethernet supports both 10Mbps and 100Mbps and can switch between the two speeds. If a sender only supports 10Mbps, it could receive and send the data whereas an 100Mbps NIC would send the data too fast (and with a different error correction system).
Question 22: What are the parts of an Ethernet frame?
Preamble: Signalizes the start of a frame.
MAC Header: Includes the sender/receiver address and marks where the frame ends.
LLC Protocol Data Unit: Contains the data.
MAC Trailer: Used for error-control.
Question 23: Why should CSMA/CD networks be built so that no more than 50 percent of their capacity is dedicated to actual network traffic?
When using CSMA/CD, the traffic should be below 50 percent because otherwise the delays when using the network will become very long because the user requests cannot be send properly.
Question 37: During the 1990s, there was intense competition between two technologies (10-Mbps Ethernet and 16-Mbps token ring) for the LAN market. Ethernet was promoted by a consortium of vendors while token ring was primarily an IBM product, even though it was standardized. Ethernet won, and no one talks about token ring anymore. Outline a number of reasons why Ethernet might have won.
First of all, the development of a consortium might be a produce that supports multiple interfaces while the IBM product represents only one kind of technique. Therefore, IBM was selling a lot of products at higher costs whereas the Ethernet technology was getting cheaper and a lot of people could afford it. This led to further price drops and led the Ethernet to a de facto standard. Second, the 16-Mbps token ring was less reliable then the Ethernet.
Question 43: You have been hired by a small company to install a simple LAN for their 18 Windows computers. Develop a simple LAN and determine the total cost.
The basic network looks like an 18 machine network connected via switch. To ensure further upgrade possibility, I would recommend to use a bigger switch then it is actually needed right now. I assume that every computer needs a network card and that we need 2 backup cards if one is going down or we need to add another PC without waiting for a long time. The length of the network cables should be around 10 meters in order to reach every workstation. Because cables can be destroyed easily, there has to be a backup of about 20% of the needed cables.
We only need transmission rates like 10/100Mbps because a small company basically does not send an amount of data that would justify a different technique with much higher costs.
Components:
24 port switch 10/100Mbps / $350.0020 NICs 10/100Mbps PCI / $200.00
22 CAT-5 Network Cables (10m) / $154.00
Total / $704.00
These costs only determine the basic hardware that is needed. Additionally, there are much higher costs for building and maintaining the whole network.
Chapter 3
Question 1: What does the transport layer do?
The Transport Layer has three major functions: First, it translates the requested server address (like a web address) into a destination IP. Second, the Transport layer splits the whole data into several smaller packages (segmentation) which can be transmitted with a higher accuracy. The used protocol is called TCP/IP. It contains data like a unique IP-Address for the sender and receiver, a sequence number and a port number to indicate for which application the data should be used. Finally, the Transport Layer checks whether the packets have reached the destination, but does not check for any problems during the way.
Question 2: What does the network layer do?
The Network Layer basically deals with routing: find a way to the next computer while taking the fastest route.
Question 3: What are the parts of TCP/IP and what do they do? Who is the primary user of TCP/IP?
It contains TCP, the Transport Layer part, which segmentates and numbers each packet. After that, this Transmission Control Protocol (TCP) reassembles the data at the receiver.
It also contains IP, the Network Player part, which is responsible for addressing and routing.
This protocol is mostly used on the internet.
Question 7: Compare and contrast the three types of addresses used in a network.
The Mac-address uniquely identifies a network component like a NIC or switch. It is an integrated number within the hardware part.
The IP-address is used to identify a computer within the network. They are unique but can be switched between several machines in order to reduce the number of addresses.
A port signalizes which application should use the outgoing/incoming data. Without assigning a port, the data could not be processed by the right program (or there can only be one service (like a telephone) where it is not necessary to identify how the data should be used)
Question 8: How is TCP different from UDP?
The User Datagram Protocol (UDP) only has a small amount of overhead added to the data that is transmitted. This includes a source port, destination port, the message length and an error control. It lacks of an error checking method like CRC.
Question 10: What is a subnet and why do networks need them?
Subnets are used to create smaller sub networks within an organization (for every department) in order to get an overview of the IP distribution. It allows an easy maintenance and provides faster transmission rates.
Question 11: What is a subnet mask?
A subnet mask determines which part of the IP-Address is local and whether the requested IP is within the same network. It hast the same structure like an IP-Address. For example, a subnet mask of 255.255.0.0 indicates that computers that have the same first 2 bytes will be in the same sub network.
Question 12: How does dynamic addressing work?
Dynamic addressing is mostly accomplished by the Dynamic Host Control Protocol (DHCP). When a computer is started, it sends a message into the network that is recognized by the DHCP server. This server distributes either a predefined or next available address to the client. This ensures that no address is used more than once in a network. In addition, in limits the number of IP-addresses needed.
Question 14: What is address resolution?
The address solution translates the address used by the application layer into a network layer address. This is necessary to send a message. TCP/IP basically uses two different techniques.
Question 15: How does TCP/IP perform address resolution for network layer addresses?
It uses a service called Domain Name Service (DNS). This service stores thousand of internet addresses and their IP addresses. The computer sends a DNS request to the DNS server and this server respond with the desired IP address.