Computer Networking: Test 3

Transcription

1 Prof. Silvia Giordano 18 June 2002 Computer Networking: Test 3 You have two hours for this exam Exercise 1 1) What is the size of the multicast address space? 2) Suppose that two different multicast groups randomly choose a multicast address. What is the probability they choose the same multicast address? 3) Suppose that 1000 multicast groups are ongoing at the same time and chose their multicast group address at random. What is the probability they interfere with each other? 4) Multicast tunneling means that IP multicast datagram are carried inside of an IP unicast datagram. How does the IP router at the end of the multicast tunnel know that the unicast datagram contains an IP multicast datagram? 5) How many multicast addresses do you need for all hosts in the network N depicted in Figure 1, assuming there is just one multicast group? LAN 2 LAN 1 G H K A B R2 D E LAN 3 C R1 to the Internet F Figure 1 Page 1

2 Solution to Exercise 1-20 points 1) 32-4 =28 bits are available for multicast addresses. Thus, the size of the multicast address space N= points 2) The probability that two groups choose the same address is 1/N=2-28 =3.73 * points 3) The probability that 1000 groups all have different addresses is N(N-1)(N-2) (N-999) = (1-1/N)(1-2/N) (1-999/N) N 1000 Ignoring cross product of terms, this is approximately equal to ( ) = 1-(999*1000)/2N= points N 4) The IP router at the end of the tunnel will use the protocol number in the "Upper Layer Protocol" field of the IP packet to determine which upper layer protocol to pass the IP packet. Thus, IP doesn't really know that the IP packet contains a multicast datagram. This is only discovered when the upper layer protocol (which will perform the multicast copy and routing) "opens" the IP datagram it is handed by the IP layer. 5 points 5) Hosts have no specific multicast address. The multicast address is unique: the one for the multicast group. 3 points Page 2

3 Exercise 2 Consider the case of a group of stations, which use CSMA/CD access protocol. Suppose at time t=0 the channel is idle, and that the channel propagation speed is v (in meters/second). From that time on only A and B are active. Let m (in meters) be the distance between A and B. S denote the slot time (in sec). At t=0, A sends a fresh frame to B. At time t=ε, ε 0 and smaller than the propagation time between A and B, station B attempts for the n th time to send a given frame to A. For simplicity, let n 6. 1) At what time do stations A and B detect the first collision? 2) Could A detect the collision if the frame transmission time is less than m/v? 3) After the collision and the mechanisms to detect the collision, the stations enter an exponential backoff phase. Call δ (in seconds) the duration of the jam signal. At what times can stations A and B start to use the channel (e.g. start to send a frame) according to CSMA/CD (so: assuming the exponential back-off mechanism)? 4) Suppose, in the previous question, that a third station C, in the exact middle of A and B (distance m/2 from A and m/2 from B), also sends a frame to A at time t=ε /2. At what time do stations A, B and C detect the first collision? Solution to Exercise 2-21 points 1) Station A will detect the first collision at time t=τ +ε and B at time t=τ, with τ =m/v 5 points 2) No, because frame transmission ends before the frame is propagated. 4 points 3) According to the exponential back-off mechanism, station A senses the channel at times t=τ +ε + δ + ks, with K=0,1. However, at time t=τ +ε + δ (k=0), the channel is still busy because of the JAM signal. It will be idle at t=2τ + δ. Thus, station A can start to use the channel at times t=2τ + δ and t=τ +ε + δ + S. The same reasoning applies to station B, can start to use the channel at times t=2τ + ε + δ and t=τ + δ + ks, with K=1,2..2 n points Page 3

4 4) Station A will detect the first collision at time t=τ/2 +ε/2; station B at time t=τ/2+ε/2, and station C at time t=τ/2, with τ =m/v 5 points Page 4

5 Exercise 3 Consider the network N in the diagram2: three LANs interconnected by routers R1, and R2. N has address /26. 1) Assign IP addresses to all of the nodes (hosts/routers) in N. 2) Define the routing table at A such that routing between LAN1 and LAN2, as well as the routing between LAN1 and LAN3 is optimized. 3) Suppose that at each host of LAN1 you want to define only the default gateway, and you must choose R1. How does A s routing table look like in this case? What happens when A opens a telnet connection with G? 4) In this case, how does A s routing table look like afterwards? At that point, what happens when A opens a telnet connection with D? 5) What happens when, after the connection with G and D, A opens a telnet connection with H? LAN 2 LAN 1 G H K A B R2 E D LAN 3 C R1 F to the Internet Figure 2 Page 5

6 Solution to Exercise 3-22 points 1) 3 LANs with different subnetting: LAN1=(A, B, C, R1, R2) and LAN2=(R2, G, H, K), and LAN3=(R2, D, E, F), taking into account that only the last 6 bits (the CIDR is /26) are available for addressing. 5 points 2) Table at A: 2 points LAN2 R2 2 LAN3 R2 2 3) Table at A: 5 points A receives an ICMP message from R1 that informs it that the shortest route to G is via R2. 4) A modifies its routing table as follows: 5 points G R2 2 A receives an ICMP message from R1 that informs it that the shortest route to D is via R2. 5) A modifies its routing table as follows: 4 points G R2 2 D R2 2 A receives an ICMP message from R1 that informs it that the shortest route to H is via R2. Page 6