1 Internet Addresses (You should read Chapter 4 in Forouzan) IP Address is 32 Bits Long Conceptually the address is the pair ( NETID, HOSTID ) Addresses are assigned by the internet company for assignment of names and numbers ICANN ICANN: Protocol Support Organization (PSO) Domain Name Supporting Organization (DNSO) Address Supporting Organization (ASO) Regional Internet Registries American Registry for Internet Numbers (ARIN) Reference “ IP Fundamentals ” Chapter 1 and 2 by Maufer

2 Formats for IP addresses: (classful addressing) Class A Network (7 bits)Host (24 bits) Network (14 bits) Network (21 bits) Host (16 bits) Host (8 bits) Multicast address (28 bits) ( There is a class E reserved for future use) Historically a class A address was assigned to networks with > 2 16 (65, 536) hosts Class B to networks with 2 8 (256) to 2 16 hosts Class C to networks with < 2 8 hosts Class B Class C Class D

3 Internet Addresses Cont. For Humans easier to deal with decimal so dotted decimal notation is used. Group 8 bits.8 bits.8 bits.8 bits and use decimal numbers Example => Class Lowest Address Highest Address A B C D Some addresses are reserved Example is reserved for loop back

4 IP version 4 uses 32 bit address billion addresses 32  and six billion live on earth at present Classless inter-domain routing (CIDR) discussed later slowed down address exhaustion. Scaling Issues: 1. Eventual exhaustion of the IPV4 address space. 2. Ability to route traffic between ever increasing number of networks that comprise the internet.

5 IP addresses cont. Class A all 0’s network number is used to represent the “ default” route ( ) this is a routing table entry which means any destination not matching any other table entry should be sent to the default route. (more later) Class A all 1’s network number is loopback usually means “ This same machine” all 127.anything.anything.anything is loopback. There are a total of = 126 class A networks. There are = 16, 382 class B networks (again all 0’s & all 1’s network addresses reserved) There are = 2,097,150 class C networks. There are 2 28 = 268, 435, 456 class D multicast addresses.

6 Notation: Class A has 8 bit network prefix this is called a slash 8’s /8 can write as 127/8 or /8 / 8 prefixes ( class A) Classfull addressing allows using only the first 1,2,3 or 4 bits to determine the class. (I.E. if MSB = 0 => class A / 8 address block has 2 31 out of 2 32 total IP addresses => / 8 takes up 50% total IPV4 addresses / 16 prefixes ( class B) 2 30 total addresses => 25% total IPV4 addresses / 24 prefixes ( class C )2 29 out of 2 32 => 12.5% total IPV4 addresses Class D with 1110 most significant bits used for multicast Class E with 1111 most significant bits and are reserved for experimental use

7 In the Beginning - IP addresses assigned based on request, not need - 32 bit address was thought to be plenty - classful addresses easy to understand and implement but not efficient / 24 supports 254 hosts too small / 16 supports 65,534 too large

8 IP Routing For a host IP routing is simple: If destination is directly connected to host then IP datagram is sent directly to the destination. otherwise host sends datagram to a default router The IP layer can be configured to act as a router in addition to acting as a host A host never forwards datagrams from one of its interfaces to another, while a router forwards datagrams IP Layer has a routing table in memory that it searches each time it receives a a datagram to send Each entry in routing table contains: - Destination IP address - IP address of a next hop router or IP address of a directly connected network - Flags (specify if network or host address) - Specification of which network interface the datagram should go to for transmission

9 IP Routing Performs The Following Actions: 1. Search routing table for entry that matches complete destination IP address 2. Search table for match of destination network IP. This check must take into account a possible “ Subnet Mask ” ( Defined Later ). 3. Search for default entry. 4. If none of the above discard and send host unreachable or network unreachable error back. Note: Hop by hop nature of this process.

10 Classical Subnetting Reference “ IP Fundamentals ” Chapter 3 Extended-Network-Prefix Network - PrefixHost - Number Network - PrefixSubnet - NumberHost - Number Before : After : Subnetting attacks the expanded routing table problem by making the subnet structure of a network invisible outside the organization’s private network. External Internet Does Not Need To Know Internal Subnet Structure

11 Example : / 16 Subnet 1 R A R C Subnet 2 R B R D Internet Subnet 5 Subnet 4 Subnet 3 R B R A Internet sees I need 5 subnets so I use 3 bits in the subnet - number field

12 This means I have Subnet Number Binary representation of least-significant two bytes Available host addresses within each subnet Need a method to know how many bits are used for subnet-number => subnet mask In this example the subnet mask is

13 Subnet Mask At boot time a machine gets its own IP address ( stored on disk for example ) Host also needs to be told how many bits are to be used for subnet ID and how many for host ID. This is the subnet mask. Subnet mask is 32 bit value containing “one bits” for the network ID and subnet ID, “zero valued bits” for host ID. Two examples for class B addresses: Class B Subnet mask = 0Xff ff ff 00 This is a /24 extended network prefix = (Subnet masks are only occasionally written in Hexadecimal) Class B 16 bits8 bits Net ID Subnet ID Host ID Net ID Subnet IDHost ID 16 bits 10 bits6 bits Subnet mask = 0Xff ff ff C0 This is a /26 extended network prefix =

14 When a host is given its own IP address and its subnetwork mask it can then figure out: Am I class A, B, or C address ? ( Look at higher order bits ) Where is the boundary between the network ID and the subnet ID ? ( Defined by class definition ) Where is the boundary between the subnet ID and the host ID ? ( Host ID is 0’s in mask )

15 Subnet Addressing Hosts are required to consider the Host ID Portion of an IP address as a Subnet ID Part and a Host ID Part. Class A and class B addresses have too many bits allocated for the host ID and respectively (-2 address because all 0’s all 1’s for a host ID not allowed ) Example: Class B 16 Bits8 Bits Net ID = Subnet ID Host ID => 256 subnets with 254 hosts per subnet (Assuming can use all 1’s and all 0’s subnet) Subnet hides the details of an internal network from external routers so that the external Router table is smaller To reach any host whose ID address begins with external routers only need to know the path to the “ Gateway” router for the entire subnetwork.

16 Gateway etc If we used multiple class C addresses external routing tables need at most one entry for each of the multiple class C addresses. Subnetting reduces the size of routing tables. Aside: A host address has a non zero ID and ID’s one host, a network address has a host ID of Zero and ID’s all hosts on that network. *

17 Example: For a class B host address of and a subnet mask of Note we have 8 bits for subnet and 8 bits for host ID If a destination IP address is bits 8 bits 8 bits Class B Subnet mask = Class B Network Subnet The network ID’s are the same ( ) the subnetwork ID’s are different ( 1 and 4 )

18 Cont If the destination IP address is the network ID’s are the same ( ) the subnet ID’s are the same (1) the host ID’s are different. If the destination IP address is ( This is a class C address ) The network ID’s are different IP routing function makes comparisons like this all the time [ REF: “TCP/IP Illustrated vol 1,” Stevens CH 3]

19 Subnet addressing Plan 1. How many total subnets does the organization need today? 2. How many total subnets will the organization need in the future? 3. How many hosts on largest subnet today? 4 How many hosts on largest subnet in future? Example Problem: An organization has been assigned the network number /24 and it needs to define six subnets. The largest subnet is required to support 25 hosts. Answer: 6 subnets => 2 3 = 8 3 bits Use a /27 extended - network - prefix => This leaves 5 bits to define host addresses on each subnet since all 0’s and all 1’s host addresses are reserved there are = 30 assignable host addresses

20 Cont This is >= 25 specified so ok Base Net = /24 Subnet # = /27 # = /27 # = /27 # =.96/27 # =.128/27 # =.160/27 # =.192/27 # =.224/27 Note: All multiples of subnet #1 (the 32)

21 All 0’s and all 1’s subnets Initially all 0’s and all 1’s subnets were forbidden. This was true to eliminate the following confusion: ( in a classful router ) When is an advertised route we do not know for example if we are advertising / bits or / bits Which is a subnet with 000 in the Extended-Network-Prefix - Without knowing the Prefix - Length or mask, a router cannot tell the difference in these two cases. ( Is it a route to an all 0’s subnet or is it a route to entire network? ) - Early routing protocols like RIP-1 do not supply a mask or Prefix - Length with each route - Now days, with the development of routing protocols that do carry this info you can use all 0’s and all 1’s subnet

22 All 0’s And All 1’s Host Addresses All 0’s host number is used to identify the base network ( or subnetwork ) All 1’s host number represents broadcast address for the network ( or subnetwork ) Example Possible Host Addresses For A Subnet Given subnet # = /27 Host # = Host # = Broadcast For This Subnet is = Aside: One can deploy network numbers from the private address space ( see RFC ) for internal connectivity and use a network address translator ( NAT ) to provide external internet access