Outline

IP Addressing
IP forwarding
ARP
IP header

IP addressing

IP address can be logically split into two parts:
- network part used for routing (locator)
- network+host part for final delivery (identifier)
- see below for details of final delivery
static subdivision -- address classes
dynamic subdivision -- subnets, CIDR

Original IP addressing

static subdivision
the address by itself is sufficient for finding network and host parts
class A: "0"/1, network/7, host/24
class B: "10"/2, network/14, host/16
class C: "110"/3, network/21, host/8
concerns:
- class C too small, many people need class B address, so inefficient use of addresses
- no address aggregation, so e.g. each networks at UH is routed to independently
all-0s and all-1s host part is reserved, cannot be used for a host

Current IP addressing

Classless IP addressing (CIDR)
network mask identifies network part of address
subnetting: many subnetworks in a single larger network (e.g. UH has 128.171/16, ICS has 128.171.10/24 and 128.171.20/24)
can give 9 bits to someone who needs 257 addresses
provides routing aggregation
provides efficient address assignment

In-class exercise

individual exercise
identify the class of the following addresses:
- 193.45.67.89
- 199.45.67.89
- 123.45.67.89
given the following subnet masks, specify the size of the subnet (maximum number of hosts) and give the broadcast address for the subnet
- 255.255.248.0
- 255.255.255.128

IP Configuration

every IP host and router is configured with:
- an IP address on each device
- a subnet mask for each device
- a "default" router (must be on one of our networks)
- possibly other routes (next hop must be directly connected to our networks)
every IP host and router must somehow be able to send to a host on an attached network knowing only its IP address:
- on Ethernet, FDDI, ATM, use ARP
- with PPP, we are told the remote address
- the reachable addresses might be configured by hand

IP Forwarding

compare the network part of the destination address to the network part of address of each of our network devices
If the destination is on an attached network, send directly to the destination
Otherwise, if we know a route to the destination, send to the next-hop router
Otherwise, send to the default router

ARP

Address Resolution Protocol
only works on network that allow broadcast (logically shared medium)
broadcast request, listen for reply, give up after a while
cache addresses from requests, replies
format:
no configuration necessary!

ARP Format

16 bit: hardware address space (e.g., Ethernet)
16 bit: protocol address (e.g., IP)
8 bit: byte length of hardware address (n)
8 bit: byte length of protocol address (m)
16 bit: opcode (REQUEST = 1, REPLY = 2)
n bytes: hardware address of sender of this packet
m bytes: protocol address of sender of this packet
n bytes: hardware address of target of this packet (if known).
m bytes: protocol address of target.

IP Header

{\tiny

    0                   1                   2                   3   
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |Version| HLen  |Type of Service|          Total Length         |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |         Identification        |Flags|      Fragment Offset    |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Time to Live |    Protocol   |         Header Checksum       |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                       Source Address                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                    Destination Address                        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                    Options                    |    Padding    |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

}

IP Header

all fields are big-endian
version number must be first field
Type of Service is defined but not widely used
Total Length is needed for protocols that may pad the packet (e.g. ethernet)
header length used for options
TTL/Time To Live (really a "maximum hop count") kills packet in case of routing errors
protocol: 1 (ICMP), 6 (TCP), 17 (UDP)

IP Header

header checksum (RFC 1071, http://www.cis.ohio-state.edu/htbin/rfc/rfc1071.html) only protects header
source address might be useful for responding
fragmentation fields allow us to send packets larger than the underlying network's MTU:
- packet ID tells us which packet this is for
- fragment offset ( * 8) is the byte offset for this fragment
- flags tell us:
  - don't fragment
  - this is the last fragment
minimum MTU for carrying IP is 576 bytes