As a french exchange student, my english is probably not

Question 1. 1. A computer network is, in my opinion, the most well known network of all, the one everyone thinks about when you evoke computer sciences.
Télécharger le PDF
174KB taille 0 téléchargements 325 vues
commentaire
Report
Principles of computer networks

Assignment 1

As a french exchange student, my english is probably not perfect. I tried to avoid a maximum of grammar mistakes but itʼs possible that a few remain. Thank you for your comprehension. Question 1 1. A computer network is, in my opinion, the most well known network of all, the one everyone thinks about when you evoke computer sciences. Itʼs also something very general. It can represent a few or millions of computers linked together by many different ways : the most common ways are with an ethernet cable (and an ethernet hub/switch), with wireless technology (WIFI) or by telecom ways, like internet does it. We usually talk about the hardware, about the physical computers, when we evoke computer networks. 2. Most of the time, the name of a network depends on the type of the informations its members exchange. For example, a network made to exchange analog voice messages is called a telephone network. A data network is just a network made to exchange digital information, such as files for example. 3. A high speed network is something really relative, like speed is. Most of the time, this term is used to designate a network where data is exchanged fast enough to able its users to use applications which require a large bandwidth. These days, this applies to video streaming (and even high definition video streaming), on-line video games, exchanging gigabytes files... 4. One of the different ways to connect the computers of a local network is to plug them all together in a same node called a switch (or a hub, but itʼs not used a lot anymore, because much slower). A few years ago, most of the switches and most of the computersʼ ethernet cards were only able to transmit and receive information with a 10 or 100 Mbps flow. A gigabyte network is a network where all the computers and the switch are equipped with a technology witch permits to exchange information with a 1 Gbps flow. 5. A P2P (peer to peer) Network is a way to connect people and let them quickly access to files without having to connect to a high speed server where the files are : the files are shared by all the hosts, and to download a file, you just pick bits and pieces of it in each personʼs computer (that person must share the file though). These persons may separately have a poor upload bandwidth, but when all put together, download becomes very fast! It has become a very interesting way to share large files without worrying about having appropriate servers to take care of it. Itʼs frequently used to share linux distributions for example, and also illegal data like movies, mp3s and video games. 6. A sensor network is represented by a few or many sensors, each equipped with a small operating system, and able to communicate together. It works in the same way

Thibault Crawley

Principles of computer networks

Assignment 1

that an ad-hoc network, meaning by that that each sensor communicates with his neighbors only, so the information has to be routed by a number of other sensors to reach its destination. These networks are usually used in specific contexts, where the need of something discrete and powerful is important. Lots of domains plan to (or already) use them : military, medical, environmental. They can provide information about natural phenomena in real time for example. 7. A social network is a way to connect people who know each-other together with for example a web site (or any other support). Every body thinks of course of the well known Facebook when the word social network comes up. 8. Ubiquitous Networks are networks which provide services for someone without this person being aware of it, and uses this service without relating it to a computer system for example, when you can easily control your whole house, like the lights, the fridge, and every thing else imaginable. 9. Wireless, in my opinion, is not really a property of the network itself, meaning that is only refers to the technology used to connect “physically” together the members of this network. You can say that a network is wireless (without wires between the hosts) if the medium used to able the hosts communicate between each-other is the air. 10. The definition is quite the same than the one above, except that the medium is a cable or a wire. Question 2 1. Milgramʼs number is the result of an experience by Stanley Milgram who tried to show that everyone on earth can be connected by 6 nodes. These nodes are others persons. It has never been theoretically proved, but this year, Microsoft researchers proved it experimentally thanks to the windows live messenger software. Its shows that the earth population is a kind of network, closer to a ad-hoc architecture : if you want to contact a certain person, you need to ask to the person you know who might be part of the chain which indirectly links you to the desired person. 2. Dunbarʼs number is the result of a research done by Robin Dunbar : it shows that a human can stay in touch with a maximum of 150 friends only by using usual ways of communication. Above this number, rules have to be created for the group to stay stable. 3. Mark Granovetter is an american sociologist. He is well known for his famous book The Strength of Weak Ties. Week ties, in this context, represent the links with the persons we donʼt know very well, person we have met a few times, friends of other friends, people with who we donʼt feel really close. It seems though that these persons can be much more useful for personal or professional evolution (finding a job, interesting contacts or relationships) than people with who you are closer, like family members or childhood friends. This is because intimate persons think they

Thibault Crawley

Principles of computer networks

Assignment 1

know you well, and that is why they will probably not be a source of evolution for you. The relation with networking is here quite the same as in the previous question, which is links between people in the world. Question 3 Voice over IP is simply a new technology which ables audio communication using the internet tool, and more specifically the TCP/IP protocol. The most well known uses of it are of course the popular softwares Skype or MSN, which give the possibility to its users to simply have audio conversations with anyone around the world. But VOIP is also used in video games (if a voice communication system between the players exists in the game). Today, when an Internet service supplier proposes free telephony, the technology used is mainly voice over IP! To be able to use it, you need a special modem, or “box”, on which you connect a classic phone. To explain how it works I will distinguish different cases : calling a VOIP phone from your VOIP phone connected to your box, calling a phone connected to the regular phone line, and finally calling another user of a VOIP software like skype. When someone speaks into his regular phone, his voice becomes an analog signal. This signal is then transformed into a digital one thanks to an integrated circuit of the box and also to a codec (or compression algorithm). The quality and the size of the digital data we now have depends on this codec. Usually, it tries to reduce a maximum the size of the message with keeping a good audio quality. Then, the box cuts the message in packets which contain about 20 ms of sound each. These packets can now be sent thru the providerʼs ADSL network. A soft switch brings the data to the right destination (things are a bit more complicated when the provider of the caller is different from the receiverʼs). Once the information arrives to the destinationʼs box, it rebuilds the digital signal and then transforms it to an analog one, so the receiver can hear it in its phone. When the destination is connected to a regular phone line, the soft switch transmits the packets to a gateway which will rebuild the signal, and transcode it with another codec, and then send it to the destination. The only difference when you use a VOIP software is that the transformation of the message into packets is made by the software itself, and that the codecs compress usually a bit more the data. The main reason why VOIP is important (or its most obvious advantage) is that you donʼt have to pay something else that your internet connection to use it! Itʼs free to call all over the word, and in an unlimited way! It is also more and more used by professionals for audio conferences for example. Voice over IP is everywhere today, and is an important part of the internet world.

Thibault Crawley

Principles of computer networks

Assignment 1

Question 4 1. Q2 :

Q3 :

Q4 :

Thibault Crawley

Principles of computer networks

Assignment 1

2. A node x1x2...xn has n neighbors. The number of digits of a node is exactly the same as the number of dimensions of the hypercube it is part of: when you add a new dimension, you link each node to one new one. Question 5 1. a) route determination: layer 3. b) flow control: layer 4. c) error correction: layer 7 or 4. d) bit to signal conversion: layer 1. 2. IP (internet protocol) is a protocol of the third layer of the OSI model, which role is to find the best way to bring a packet to its destination. Unlike TCP, it does not assure that the data arrives to destination. This is called “best effort” practice. 3. TCP (Transmission Control Protocol) in responsible for exchanging informations between software applications (like emails, files..) : it works at a higher level than IP, and unlike IP, it checks if the data arrived correctly. Because IP doesnʼt handle receiving verification, TCP handles all these problem for it. TCP was basically created so as the software developers donʼt have to take care of low level networking details. Question 6 1. On a 8 Gbps link, 8 Giga bits are 1 second wide. You just have to divide this second by 8 000 000 000 to know how wide is a bit : 1 / 8 000 000 000 = 1.25e-10 seconds. 2. To have the length in meters of a 1e-9 sec wide bit in copper wire when the speed of propagation is 2.3e8 m/s, you just have to multiply 1e-09 by the speed of propagation : 1e-09 x 2.3e8 = 0.23 meters. Then you multiply this result by 100 to have the length in centimeters. The final result is : 23 centimeters. Question 7 1. Letʼs say that our packetʼs size is x bits. The linkʼs speed is 100 Mbps. It takes x / 100 000 000 seconds to put all the bits of our packet on the medium. Letʼs call this duration d1. After d1, the last bit of our packet is on the medium. It then takes 30 micro seconds for this last bit to arrive at the switch S. When that bit arrived, the whole packet is there.

Thibault Crawley

Principles of computer networks

Assignment 1

The same operation is repeated 50 micro seconds after : we have another d1 duration for all the bits to be on the medium, then we wait 40 micro seconds for the last bit to arrive at the host B. It finally took d1x2 + 30 micro seconds + 50 micro seconds + 40 micro seconds = 120 micro seconds + 2d1 to transmit a packet from host A to host B. 2. The RTT of a packet is the delay for that packet to reach its destination added to the delay for the source to receive the acknowledgment sent back by the destination. If the acknowledgment packageʼs size is the same that the received package, then the TTL is simply the double, i.e 2x(120 micro seconds + 2d1). If itʼs not (lets say its size is y bits), we will have to bring up a d2 = y / 100 000 000 seconds and calculate with the same method the delay of transmitting the acknowledgment package for B to A. The TTL will be in this case : (120 micro seconds + 2d1) + (120 micro seconds + 2d2). 3. To transmit k packets from host A to host B (if these packets are all the same size), the total time will be k x (120 micro seconds + 2d3), with d3 = z / 100 000 000 and z : number of bits in a packet. Question 8 1. One frame contains 360 x 240 = 86 400 pixels. Each pixel is 12 bytes, or 12x8 = 96 bits (1 byte = 8 bits). We have 86 400 x 96 = 8 294 400 bits by frame. We have 40 frames every second, so 8 294 400 x 40 = 331 776 000 bits are transmitted in this transmission technology. 2. One frame contains 640 x 480 = 307 200 pixels. Each pixel is 16 bytes, or 16x8 = 128 bits (1 byte = 8 bits). We have 307 200 x 128 = 39 321 600 bits by frame. We have 50 frames every second, so 39 321 600 x 50 = 1 966 080 000 bits are transmitted in this transmission technology. Question 9 Lets see the different steps followed by the request, and the response : The request: The request is first transmitted to the satellite 1 witch is at 40 000 kms. It takes 40 000 / 300 000 = 0.1333 seconds to reach that satellite. To travel from the satellite 1 to the satellite 2, it takes 100 000 / 300 000 = 0.3 seconds (there are 100 000 kms between both satellites) . Finally, it takes another 0.1333 seconds to reach the ground. So for the request to reach the server, it takes 0.1333x2 + 0.3 = 0.5666 seconds. The response: It theoretically takes the same time for the response to travel from the server to the client, the message just follows the same path backwards. The best case propagation

Thibault Crawley

Principles of computer networks

Assignment 1

delay in response to a request is 0.5666x2=1,1222 seconds, if you consider of course that the server treats the demand instantaneously. Question 10

For the manchester signal, I searched on the internet for more precision, and it seemed like the version of Manchester that is mostly used is the Differential Manchester encoding, so that is the one I represented here. This is how it works : At the beginning of a bit, a “1” is coded by a transition, and a “0” by an absence of transition. To be able to do these transitions at each beginning of bit, an other transition has to take place in the middle of a bit time. Explanations for the Differential Manchester encoding were found here (in french) : http://pagesperso-orange.fr/arsene.perez-mas/transmission/seriel_sync/manchester.htm Question 11 1. A cycle is 360 degrees, or 2 x pi radians. So the phase of a sine wave which is offset one-sixth of a cycle with respect to time zero is 360/6 = 60 degrees or (2 x pi)/6 = pi/3 = 1.05 radians. 2. One of the well known properties of the trigonometric circle is that sin(x + pi/2) = cos(x). We also have cos(x + pi/2) = -sin(x). So the offset of a cosine wave with respect to a sine wave is pi/2. Question 12 1. 1 ms = 1000 microseconds so 100 ms = 100x1000 microseconds = 100 000 microseconds. 2. 100 ms = 0,1 second. f=1/t, so f=1/0,1 = 10hz = 0,01 Khz.

Thibault Crawley