As network admins, we spend a lot of time
concerning ourselves with IP addresses –
assigning them,
filtering them, etc. we don’t think about MAC
addresses that often, but data from Host A to Host
B must have a destination IP address for Host B
and a destination MAC address as well.
Well, this can be learnt from a DNS server which
interprets the names of computers on the network.
The natural question is “ How does Host A know the
IP address of the DNS Server in the first place?”
That happens in one of two ways:
The DNS server address is hard-coded on Host A
The DNS server address was learned via DHCP
We’re going to take a look at DHCP later in this
section, but here’s the partial output of ipconfig /all
on a windows PC using DHCP. The DNS server IP
addresses are In bold.
You’ll see the bold part I highlighted there. That is
the DNS server locations. That’s how this PC
knows where to send the DNS request in the first
place. One thing we don’t see there is the ARP
server. ARP is the Address Resolution Protocol and
that’s the protocol that allows a device to get a
layer 2 address ( MAC Address). There’s no such
thing as the ARP server that’s why we don’t have
one.
But Host A need to have the MAC address to send
data successfully to Host B. well since there’s no
ARP server, the ARP process uses a series of
broadcasts and replies.
Host A is the host sending a MAC address of a
remote device, so it’ll be Host A that sends out the
initial ARP Request. This request is a Layer 2
broadcast, meaning…..
The source MAC address will be that of Host A
The destination MAC address will be ff-ff-ff-ff-ff-ff
(layer 2 broadcast adress)
The source IP address will be that of Host A
The destination IP address will be that of Host B
(learned via DNS)
It is very important to remember that the ARP
request is a broadcast and the ARP reply is a
unicast. So now Host A has all the data it needs
about Host B and it can successfully send data to
host B.
source: http://tutorialslodge.com/arp/