ps3.typ (11040B)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 | #show link: underline == Problem Set 3 #h(1fr) Jack Sangdahl === Problem 1 Figure 1 below is a typical home network setup. An ISP modem provides internet service; a wireless router is connected to the ISP Modem via Ethernet. Hosts A, B, and C are connected to the wireless router to access the internet. #align(center)[ #image("img/fig1.png", width: 70%) Figure 1 ] + In order for hosts A, B, and C to access the web server, Network Address Translation (NAT) with random port mapping needs to be enabled for both the wireless router and the ISP modem. Assume hosts will pick a random port number between 8000 and 9000, the wireless router can choose a random port number between 2000 and 2500, and the ISP modem can choose a random port number between 3000 and 4000. Fill in the NAT table for the wireless router and ISP modem. #align( center, )[ NAT Table of Wireless Router #table( columns: 2, table.header[*LAN Side*][*WAN Side*], [`192.168.0.5:8001`], [`10.0.1.5:2001`], [`192.168.0.11:8002`], [`10.0.1.5:2002`], [`192.168.0.22:8003`], [`10.0.1.5:2003`], ) NAT Table of ISP Modem #table( columns: 2, table.header[*LAN Side*][*WAN Side*], [`10.0.1.5:2001`], [`96.120.12.18:3001`], [`10.0.1.5:2002`], [`96.120.12.18:3002`], [`10.0.1.5:2003`], [`96.120.12.18:3003`], ) ] + Now we look into the details about how packets are exchanged between host B and the web server. Assume host B sends an HTTP request packet to the web server. And the web server then sends HTTP content back to host B. Fill in the tables below to show how the packet's IP header changes along the route. #align( center, )[ Request Before Entering Router #table( columns: 2, [*Source IP*], [`192.168.0.11`], [*Source Port*], [`8002`], [*Destination IP*], [`216.58.217.36`], [*Destination Port*], [`80`], ) Request After Exiting Router #table( columns: 2, [*Source IP*], [`10.0.1.5`], [*Source Port*], [`2002`], [*Destination IP*], [`216.58.217.36`], [*Destination Port*], [`80`], ) Request After Exiting Modem #table( columns: 2, [*Source IP*], [`96.120.12.18`], [*Source Port*], [`3002`], [*Destination IP*], [`216.58.217.36`], [*Destination Port*], [`80`], ) Response Before Entering Modem #table( columns: 2, [*Source IP*], [`216.58.217.36`], [*Source Port*], [`80`], [*Destination IP*], [`96.120.12.18`], [*Destination Port*], [`3002`], ) Response After Exiting Modem #table( columns: 2, [*Source IP*], [`216.58.217.36`], [*Source Port*], [`80`], [*Destination IP*], [`10.0.1.5`], [*Destination Port*], [`2002`], ) Response After Exiting Router #table( columns: 2, [*Source IP*], [`216.58.217.36`], [*Source Port*], [`80`], [*Destination IP*], [`192.168.0.11`], [*Destination Port*], [`8002`], ) ] + Suppose now host A also runs a web server on port 8888 and it is attached to a domain name #link("http://www.mylocalhomeserver.com"). Explain what NAT entries should be added so that people from the internet can assess this web server via URL. You can assume that the above domain name is registered properly. The following entries should be added for the router and modem. IP address `192.168.0.5` should have port `8888`, and address `10.0.1.5` can have a random port number. #align( center, )[ NAT Table of Wireless Router #table( columns: 2, table.header[*LAN Side*][*WAN Side*], [`192.168.0.5:8888`], [`10.0.1.5:8888`], ) NAT Table of ISP Modem #table( columns: 2, table.header[*LAN Side*][*WAN Side*], [`10.0.1.5:8888`], [`96.120.12.18:8888`], ) ] + The wireless link at the last mile is very error prone and you would like to improve the performance. What would you do in this case? // TODO === Problem 2 Suppose a router has three input flows and one output flow. It receives the packets listed in Table 1 below all at about the same time, in the order listed, during a period in which the output port is busy, but all queues are otherwise empty. Give the order in which the packets are transmitted, assuming... #align( center, )[ #table( columns: 3, table.header[*Packet*][*Size*][*Flow*], [1], [200], [1], [2], [200], [1], [3], [160], [2], [4], [200], [2], [5], [160], [2], [6], [210], [3], [7], [120], [3], [8], [90], [3], ) Table 1 ] + Fair Queueing #align( center, )[ #table( columns: 4, table.header[*Packet*][*Size*][*Flow*][*Finish Time*], [1], [200], [1], [200], [2], [200], [1], [400], [3], [160], [2], [160], [4], [200], [2], [300], [5], [160], [2], [460], [6], [210], [3], [210], [7], [120], [3], [350], [8], [90], [3], [440], ) ] The order of packet transmission is determined by the finishing time. So as per above, packets would transmit in the following order: *3, 1, 6, 4, 7, 2, 8, 5*. + Weighted fair queueing with flow 2 having twice as much share as flow 1, and flow 3 having 1.5 times as much share as flow 1. Note that ties are to be solved in the order of flow 1, 2, and 3. #align( center, )[ #table( columns: 4, table.header[*Packet*][*Size*][*Flow*][*Finish Time (weighted)*], [1], [200], [1], [100], [2], [200], [1], [200], [3], [160], [2], [40], [4], [200], [2], [75], [5], [160], [2], [115], [6], [210], [3], [70], [7], [120], [3], [116.67], [8], [90], [3], [146.67], ) ] The order of packet transmission is determined by the finishing time. So as per above, packets would transmit in the following order: *3, 6, 4, 1, 5, 7, 8, 2*. === Problem 3 Assume TCP Reno is the protocol experiencing the behaviour shown in Figure 2. #align(center)[ #image("img/fig2.png", width: 70%) Figure 2 ] + Identify the RTT rounds when TCP runs Slow Start. (eg. From the 1st round to which?) Slow Start occurs for the initial 6 RTT's, and the final 4 RTT's. + Identify the RTT rounds when TCP runs Congestion Avoidance. Congestion Avoidance occurs from the 6th transmission, to the 19th transmission. + After the 14th RTT round, is segment loss detected by a triple duplicate ACK or by a timeout and why? Segment loss is detected by a triple duplicate ACK, as otherwise the window will have dropped to 0 + During which RTT round is the 170th segment sent? The 170th segment is sent during the 10th transmission. + Assuming a packet loss is detected after the 23rd RTT round by the receipt of the triple duplicate ACK's, what will be the value of the congestion window? The value of the congestion window will be *4*. === Problem 4 Figure 3 below shows how two disconnected LANs are connected by IP tunneling (the dashed line). For each interface, the IP and MAC addresses are shown in the figure (HW1 through HW14 are used to represent the hardware address). #align(center)[ #image("img/fig3.png", width: 70%) Figure 3 ] Now host B sends a packet to host A. Show how the packet travels along the route, describe header information along the route (B #sym.arrow.r R6, R6 #sym.arrow.r R3, R3 #sym.arrow.r R2, R2 #sym.arrow.r R1, R1 #sym.arrow.r A). #align( center, )[ Host B to Router R6 #table( columns: 2, [*Source IP*], [`125.0.2.200`], [*Source MAC*], [`HW14`], [*Destination IP*], [`125.0.2.5`], [*Destination MAC*], [`HW13`], ) Router R6 to Router R3 Inner IP header: Same as Host B to Router R6 Outer IP header: #table( columns: 2, [*Source IP*], [`23.30.1.108`], [*Source MAC*], [`HW12`], [*Destination IP*], [`128.38.81.4`], [*Destination MAC*], [`HW7`], ) Router R3 to Router R2 Inner IP header: Same as Host A to Router R1 Outer IP header: #table( columns: 2, [*Source IP*], [`23.30.1.108`], [*Source MAC*], [`HW6`], [*Destination IP*], [`128.38.81.4`], [*Destination MAC*], [`HW5`], ) Router R2 to Router R1 Inner IP header: Same as Host A to Router R1 Outer IP header: #table( columns: 2, [*Source IP*], [`23.30.1.108`], [*Source MAC*], [`HW4`], [*Destination IP*], [`128.38.81.4`], [*Destination MAC*], [`HW3`], ) Router R1 to Host A #table( columns: 2, [*Source IP*], [`125.0.2.200`], [*Source MAC*], [`HW2`], [*Destination IP*], [`125.0.2.5`], [*Destination MAC*], [`HW1`], ) ] === Problem 5 Derive the expected throughput of the following TCP congestion algorithm: The additive increment factor is #sym.alpha. Multiplicative decrease #sym.beta, which means after loss, the window's size will change from $W$ to $(1 - beta)W$. Order the throughput for each flow. `AIMD(a, b)` means the CWND increases per each round trip time and the CWND set to $(1 - beta)W$ for $W$ when the loss happens. #align( center, )[ #table( columns: 4, table.header[*Flow*][*AIMD*][*RTT*][*Loss*], [1], [$a = 1, b = 0.5$], [10ms], [$10^(-6)$], [2], [$a = 2, b = 0.2$], [100ms], [$10^(-8)$], [3], [$a = 5, b = 0.8$], [300ms], [$10^(-9)$], [4], [$a = 8, b = 0.4$], [1000ms], [$10^(-4)$], [5], [$a = 6, b = 0.5$], [100ms], [$10^(-10)$], ) ] To derive the expected throughput, use the Mathis equation for TCP throughput. $ Tau approx (alpha) / ("RTT" times sqrt(p) times sqrt(beta)) $ Where #sym.alpha is the additive increase factor, #sym.beta is the multiplicative decrease factor, $p$ is the packet loss probability, and RTT is the round trip time. Apply this formula for each flow. + $alpha = 1, "RTT" = 10 "ms", p = 10^(-6), beta = 0.5$ $ Tau_1 approx (1) / (0.01 times sqrt(10^(-6)) times sqrt(0.5)) approx 141.4 "Mbps" $ + $alpha = 2, "RTT" = 100 "ms", p = 10^(-8), beta = 0.2$ $ Tau_2 approx (2) / (0.1 times sqrt(10^(-8)) times sqrt(0.2)) approx 447.5 "Mbps" $ + $alpha = 5, "RTT" = 300 "ms", p = 10^(-9), beta = 0.8$ $ Tau_3 approx (5) / (0.3 times sqrt(10^(-9)) times sqrt(0.8)) approx 186.4 "Mbps" $ + $alpha = 8, "RTT" = 1000 "ms", p = 10^(-4), beta = 0.4$ $ Tau_4 approx (8) / (1 times sqrt(10^(-4)) times sqrt(0.4)) approx 1266.5 "Kbps" $ + $alpha = 6, "RTT" = 100 "ms", p = 10^(-10), beta = 0.5$ $ Tau_5 approx (6) / (0.1 times sqrt(10^(-10)) times sqrt(0.5)) approx 848.5 "Mbps" $ Thus, the ordered throughput for each flow is given in descending size: *5, 2, 3, 1, 4*. === Problem 6 Suppose that TCP uses the combination of quick acknowledgements (quick ACK) and delayed acknowledgements (delayed ACK). The quick ACK only triggers up to 8 packets (the CWND at the sender becomes 16 after receiving 8 quick ACKs) starting from 1 packet during Slow Start. The maximum capacity of the link is 5000 KBps, the RTT is 10ms, and 1 MSS is equal to 1KB. + About what is CWND at the time of the first packet loss? After the quick ACK phase, TCP switches to delayed ACKs, meaning the receiver sends one ACK for every two packets. In Slow Start, the CWND roughly doubles with each RTT, and begins at 16 in our current scenario. So after 1 RTT, CWND is 32 MSS, after 2 RTTs, CWND is 64 MSS, and so on. When considering packet loss, it will occur when the CWND reaches the capacity of the link, which is 5000 KBps. This would be reached after roughly 7 RTTS, or when the CWND is approximately 1024 MSS (1 MB). + About how long until the sender discovers the first loss? As found above, it takes 7 RTTs to grow the CWND to 1024 MSS, which is the point where packet loss is expected. The RTT is 10 ms, so the time to discover packet loss is given by $7 "RTTs" * 10 "ms" = 70 "ms"$. |