Review of last class

III. Layer 3 Technology
     A. Network layer characteristics
     B. Network Addressing
     C. Routing
     D. Fragmentation
     E. IP Protocol
        1. Addressing - doing over today
        2. ARP
        3. BOOTP/DHCP
        4. Intro to DNS

Administrative

1. Computer Lab Update
2. Project Topics Due Today
3. Homework returned
4. Readings/weblinks
   a. http://www.ccci.com/product/papers/pete/papers/bgp1.htm
   b. http://www.lantimes.com/handson/97/706a107a.html
   c. http://www.stardust.com/ipv6/documents/v6back.htm
      Note: read at least Part I - The Business Case for IPv6
   d. http://www.FreeSoft.org/CIE/Topics/81.htm
   e. http://www.FreeSoft.org/CIE/Topics/53.htm

Layer 3 Continued

2. IP Addressing Revisited

   Each interface on a TCP/IP internet is assigned a unique 32 bit internet
   protocol (IP) address, which is used for communications across the entire
   internetwork.  Some IP addresses are used for special delivery services
   and some are reserved for specific uses.

   Typically, IP addresses are a "pair", consisting of a network id and a
   host id. Traditionally each physical network segment receives it's own IP
   network id.

   <picture and description of router port attached to Ethernet segments>

   In theory, with 32 bits for addressing, there are over 4 billion unique
   addresses available.  However, in the Internet and in practice, much less
   may actually be available.  More on this later.


           0  1234567 8      16      24     31
           +-+-------+------------------------+
   Class A |0| netid +        hostid          | 2^7 networks, 2^24 hosts
           +-+-------+------------------------+
           +-+-+--------------+---------------+
   Class B |1|0|     netid    |     hostid    | 2^14 networks, 2^16 hosts
           +-+-+--------------+---------------+
           +-+-+-+-------------------+--------+
   Class C |1|1|0|       netid       | hostid | 2^21 networks, 2^8 hosts
           +-+-+-+-------------------+--------+
           +-+-+-+-+--------------------------+
   Class D |1|1|1|0|     multicast address    | 2^28 addresses
           +-+-+-+-+--------------------------+
           +-+-+-+-+-+------------------------+
   Class E |1|1|1|1|0|   reserved for future  | 2^27 addresses
           +-+-+-+-+-+------------------------+

   o  Class A networks can be determined by the first high order bit set to
      0.  This allows for <128 networks, but for more than 4 million hosts
      on each network.

   o  Class B networks can be determined by the first two high order bits
      set to 1 and 0 respectively.  This allows networks which need less
      than approximately 64,000 hosts.

   o  Class C networks can be determined by the first three high order bits
      set to 1, 1 and 0 respectively.  This allows for a large number of
      networks, but for less than 256 hosts per network.

   o  Class D addresses can be determined by the first four high order bits
      set to 1, 1, 1 and 0 respectively.  There is no separation of host
      and network pair.  This is a special address range for multicast
      services.  A multicast group address follows the first 4 bits.

   o  Class E addresses can be determined by the first five high order bits
      set to 1, 1, 1, 1 and 0 respectively.  This address range has been
      reserved for testing and future use.

   A. Classes and Dotted Decimal Notation

      The three primary classes are easy to work with because the network
      id and host id are split on octet boundaries.  For example, with a
      class A network, the last three octets (bytes) correspond to the host
      id.

      The value of the first octet determines the class with which the
      address falls.

      As in:  Class        Range of values
              -----       -----------------
                A            0 through 127
                B          128 through 191
                C          192 through 223
                D          224 through 239
                E          240 through 255

   B. Special Addresses

      1. Network Address

         Internet addresses can be used to refer to networks as well as
         individual hosts.  By convention, an address that has all bits of
         the host id equal to 0 is reserved to refer to the network.

         i.e. 129.65.0.0 refers to the class B network with prefix 129.65

      2. Directed Broadcast Address
 
         In the case where you'd want to send a IP packet to all hosts on a
         particular IP network, you can send a packet with all host id bits
         set to 1 while using the specific network id.

         i.e. 129.65.255.255 refers to all hosts on network 129.65.0.0

      3. Limited Broadcast Address

         Like the directed broadcast, the limited broadcast address is used
         to send a packet to all hosts on a specific network.  However, the
         limited braodcast address has all network id and host id bits set
         to 1.  This is used by stations on a IP network to send a packet
         to all hosts on it's network, without knowing the network id.  For
         example, in the case of a host booting up and before it has learned
         it's network number.

         i.e. 255.255.255.255 is a limited broadcast on the network with
              which it was sent

      4. This Computer and This Network Address

         When a address has all network id bits set to 0, it literally means
         "this network".  As does the host id when all its bits are set to
         zero, it litterally means "this" host.  This is often used as a
         source address when a station does not yet know it's IP address as
         in the case of a boot-up sequence.

         i.e. 0.0.0.0 specifies this host and this network

      5. Loopback Address

         IP preserves the class A network address of 127.0.0.0 as a loopback
         address.  All host id bits are irrelevant.  Anytime a packet is
         sent to this address, it is "looped back" through software, the IP
         protocol stack on the same computer.  This is very useful in
         testing for programmers of TCP/IP based software.

         For example, you could start a TELNET session to the local computer
         by specifying the loopback address as the destination address.

   C. Subnetting

      In the original class hierarchy of addressing, there was one problem
      which only became apparent after time.  Growth.

      Three individual issues arose.

      i.   The immense administrative overhead in managing address space.

      ii.  Routing tables in Internet routings became very large.

      iii. Address space was going to eventually be exhausted.

      To appropriately size local IP networks, subnetting is used to split
      the host id into a local network part and smaller host id part.

      <example of subnetting a class B into multiple class C sized
       addresses)

                              H1 (128.10.1.1)    H2 (128.10.1.2)
                                     |                  |
                                -+---+------------------+-------
                                 |    Network 128.10.1.0
      Internet-----[local router]+
                 ^               |    Network 128.10.2.0
                 |              -+---+------------------+-------
                 |                    |                  |
                 |            H3 (128.10.2.1)    H4 (128.10.2.2)
                 |
           All traffic to
             128.10.0.0

      The result is a form of hierarchical addressing that leads to the
      corresponding hiearchical routing.

      Not all IP networks could be as easily subnetted as the example above.
      Flexibility in subnetting allows for the local portion to be split
      between a local subnet portion and local host id portion of varying
      length, no necessarily along octet boundaries.

      For example, above we may choose to use the first 3 bits in the local
      portion for the subnetwork and the remaining 13 bits for host ids.
      This would allow for 8 subnetworks with up to 8192 hosts per subnet.

   D. Subnet Masks

      To implement subnetting, you must use a 32 bit subnet mask for each
      subnetwork.  Bits in the subnet mask are set to 1 if the network
      treats the corresponding bit in the IP address as part of the network
      address and 0 if it treats the bit as part of the host id.

      Ex: 11111111 11111111 11111111 00000000

      You may often see addresses in the form: 128.10.1.1/24.  The '/24'
      means a 24 contiguous bit subnet mask as in the example immediately
      above.

      In addition, you often specify subnet masks in sofware by using the
      dotted decimal notation which is easier for humans to read.  The
      example above would have a subnet mask of 255.255.255.0.

<Let's DO some networking!>