Small recall on the limitations of IPv4.
- The number of addresses is limited has 232 Adresses minus
the addresses of networks and broadcast.
- the tables of routing are full and it is impossible to make configurations
more reliable with the only current protocols.
Easy ways :
NAT: To limit the number of IPv4 used on the network, the NAT (Network Adress Translation)
was invented it is a method which allows "hidding" a network of the rest of the world the outgoing
IP packet on a NAT router carries all the same address what allows has companies not to use
that only one address IP known as Publique, the router keeping a table of correspondence to
know has which host to send the response of a request. One calls also this kind of method
Masquerading but this name tends has to disappear bus being only used by linux (this
process initially was indeed related to the Linux Systems) .
But this easy way blocks many applications not being conceived for.
CIDR: much among us knows the classes of addresses of IPv4 which allow a
certain distribution of blocks of addresses.
| Adress Class |
Bit of strong weight |
Adress Bloc |
Network id |
Machine id |
Number of Network |
Number of Machine/Network |
|
A |
0000 |
1-126.x.y.z |
1-126 |
x.y.z |
126 |
16 777 214 |
| B |
0010 |
128-191.x.y.z |
128-191 |
y.z |
16 384 |
65 534 |
| C |
0110 |
192-223.x.y.z |
192-223 |
z |
2 097 152 |
254 |
| D |
1110 |
224-239.x.y.z |
- |
- |
- |
- |
| E |
1111 |
240-247.x.y.z |
240-247 |
|
|
|
This way of making is from now on obsolete and is replaced by the CIDR
(Classless Inter Domain Routing) which makes it possible to optimize the size
of the tables of routings used by the routers of the large world dorsals.
The CIDR makes it possible to gather several roads in only one or contrary
to dividing networks into sub-networks according to the mask of under network
The rapid exhaustion of the addresses of class B showed that 3 classes of
primary addresses were not sufficient class A being too large C too small,
the class B being too important for many network but it was best the solution.
a more flexible manner to organize the portions identifying the network and
the machine in an address is to use the binary masks, the mask binary function
in this way: if a bit is positioned in the mask, the equivalent bit in the address
is interpreted as being a bit network
Example : 192.168.1.0/16 is under network of the old Class B with a
mask of 255.255.0.0
But : 192.168.1.0/24 is under under
customized network that the old notation of classes did not allow normally
because a segment of class B is used with a mask of class C
(WELL ON this example addresses of Privée class one uses should not hold
account of it)
To use a mask to create networks larger than the natural mask is called
suppernetting and the use of a mask instead of the class of address to determine
the network of destination is called CIDR (Classless Inter Domain Routing).
| 