Help on IP address?

Question

What aspects of IP address makes it necssary to have one address per network interface , rather than just one per host? And why does IP address tolerate point to point interface that have non-unique address or no address?

Answer 1

To put it simply, every interface MUST have a unique address. It can't work any other way. If a host has a single interface, then that address is also the host's address. If it has multiple interfaces then ALL of the addresses are addresses of the host.

It's impossible for two devices to have the same IP address and have them work. And a device that has no IP address can't work either.

Not sure where you're getting your information, but it's just not correct. I'd suggest a basic course in networking to help clear this up in your mind.

Answer 2

The IP address maps to a physical device's MAC address (6 digit hex number). The MAC address is like the serial number of the Ethernet chip, and is known as the level-2 address (physical).

The IP address (level-3) is a virtual address which you have control over (assignment), to fit as part of a network. It must be attached to one physical address. For example, use ARP (Dos/Command window or other ARP tool), and it will reveal the MAC address bound to each IP.

The OSI model specifies layers in the network connectivity model, and this answer applies to only to layers 2 (MAC) and layer 3 (Network).

For the Ethernet chip to do its job (listen on network), it must have it's own MAC address; so that the chip upon opening the packet of data, will be able to recognize that it (the packet) is meant for that interface.

By assigning an IP (routable protocol) address to the interface, you can create networks of computers on specific networks, minimizing the amount of broadcasts on the network.

To my knowledge, there is not one Ethernet chip which can have many virtual (IP) interfaces attached to it.

Point-to-point is achieved exactly because IP is a routable protocol; IP addy 1.1.1.1 maps to MAC 00-05-02-E2-A3-68, and IP 1.1.1.2 can map to any next MAC addy. Every MAC addy is supposed to be unique, which includes the manufacturer portion and serial number of the Ethernet interface.

An interface with no MAC address is supposed to obtain an MAC address when the system housing it boots up.

Hope this helps.

Answer 3

Because an IP address is x.x.x.x, and x can be between 1 to 255 only. So we're running out of addresses in the world. So we need "group addresses"

As for your second question, they probably use NAT
http://en.wikipedia.org/wiki/Network_address_translation.

Check it out.