IP address redesigned?

Hi again,
I have another question from class (which btw.. i'm not oding so great in )

Could IP be redesigned to eliminate address classes completely if each address occupied 64 bits instead of 32 bits? Explain.


I've been trying to figure out my hw questions on my own but my answers arent completely right, hence the not doing well part.
if anyone could please help me out,
thanks
felicia

 

Well, current IP (IP v4) uses 4 octets of bits separated by dots:
So 192.168.10.15 is actually 11000000.10101000.00001010.00001111 in binary. So that's where you get 32 bits.

Now, an IP address by itself isn't that useful. You need to know what part of the address is the network's address, and which part is the host's address. For that, we use the subnet mask. In classful notation, the subnet mask is either 255.0.0.0, 255.255.0.0, or 255.255.255.0. Again, we have 4 octets of bits, making another 32 bits. The problem here is that you now need to keep track of the subnet mask in addition to the IP address. Rather than having two separate addresses, you could do the following.

Place the 4 subnet address octets immediately after the 4 IP address octets. So out above host with an IP of 192.168.10.15 and a subnet mask of 255.255.0.0 suddenly get an address of 192.168.10.15.255.255.0.0 . That's 32 + 32 = 64 bits. Then the protocol would simple specify the first 4 octets as address bits, and the last 4 octets as subnet mask bits.

In practice, that wouldn't be the most efficient way to do it. Since subnet masks are almost always made to be a set number of consecutive 1's followed by all 0's (that is, 11111111.11111111.00000000.00000000), you could actually create a single fifth octet that represents the number of bits that are 1's in the subnet mask. So since 255.255.0.0 has 16 positive bits, the address could be 192.168.10.15.16.

Cisco likes to use a simlilar notation for subnetting. They will often say a host is 192.168.10.15/16. The slach separates the IP address from the subnet mask, and the 16 represents 16 consecutive 1's in the subnet mask.

IPv4 (the current one) is a 32 bit scope. IPv6 (IPng) which is currently in testing is a 128 bit addressing scope. That means 32 octets. IIRC, that is a large enough address space for roughly 6 * 10^24 addresses per square meter of the earth surface.