Exercise 9  (12. -16.4.) (Tanenbaum pp. 374 -424)

1. Answer shortly the following questions ("review type" problem: answers are usually  found  in
     lecture notes or in Tanenbaum's book).
     a) Why  is  flow control much easier than  congestion control?
     b) Buffers of a router are overflown. Which packets  will be  destroyed, the arriving ones or
          some others?
     c) What is  tunneling?
     d) Why  a firewall has both routers and  a gateway?
     e) Both IP  and ATM protocols  count checksums  only for the header?  Why for the  header and
         why only for it?
     f) Is  fragmentation needed  in concatenated virtual circuit internets, or only in datagram systems?
     g) Why are ARP and RARP protocols needed? What  is common to both these protocols?
         How do they differ from each other?

2. a) What kind of errors the network layer  has to prepare  for?  Is it possible to  have  the  data link
         layer or   transport layer to take care of these errors?
    b) What all things have to be negotiated  when establishing a virtual  circuit across many networks?
         Explain in each case what happens when wishes/ requirements are not  compatible?

3.  A  PC is connected to a 6 Mbps  network, where  token bucket  method  is used  for  congestion
     control.  Tokens are generated at  a rate of  1 Mbps. The token bucket is initially filled with to
     maximum capasitity of 8 megabits. How long  can the computer transmit at the full 6 Mbps?
     (The answer is not 1.33 s!)
.
4. a) Tunneling through a concatenated  virtual circuit subnet is straightforward: the multiprotocol
        router at one end just sets up a virtual circuit to the other end and passes packets through it.
       Can tunneling also be used in datagram subnets?  If so, how?
    b) A person with a portable computer travels to Berkeley and  to his surprise finds out that
        there is a  wireless LAN. So he doesn't have to plug his computer  in, but can he use it?
        If he can, how is it done?
   c) In Berkeley there is also a GSM network. Supposing the local operator  has a roaming contract
       with some  finnish  operator (calls  from one operator are transferred  to another). Explain  how
       a travelling  person can read his email from Berkeley,  or is it not possible?

5.  a) Convert the IP address whose hexadecimal representation is   C22F1582 to dotted decimal
         notation.
     b)  A class B network on the Internet has a subnet mask 255.255.240.0. What is the maximum
          number of hosts per subnet?

6. a) Describe,  in  suitable detail,  how IP fragments are  reassembled at the destination.
    b) Some fragments can get lost in the transmission. For this reason,  the  reassembly of fragments
        uses  a timer.  In case of timeout  the unfinished  packet is destroyed. Suppose a datagram is
        fragmented into four fragments. The first three fragments arrive, but the last one is delayed.
        Eventually the timer goes off and the three fragments in the receiver's memory are discarded.
        A little later, the last fragments arrives. What should be done with it?
    c) An IP datagram  using  the  Strict source routing  option. has to be fragmented. Do you think
        the option has to be copied  into each fragment, or is it sufficient to just put it in the first
        fragment? Explain your answer.