TCP/IP

A communication protocol is a set of implemented regulations to achieve information transfer from one point to another. There are a variety of protocols, but the most widely accepted today is the so-called TCP / IP (Transport Control Protocol / Internet Protocol) better known as IP address.

Eguibar Information Technology S.L. can help you in choosing the right strategy of implementing the IP Address space, designing the IP routing, logical IP address segmentation, the correct configuration of network devices, etc.

IPv4

Version 4 of TCP / IP (IPv4) has more than 4 billion addresses (2³²), which are divided into classes (type A, B, C …). There are Public and Private type. Even when it seems an endless number of IP, are not sufficient for the whole world (web pages, communications devices, mobile devices, virtualization, inefficient use etc.), But fortunately there are different configurations, technologies and services that help us fit address in this dispute.

IPv6

Is the evolution of IP version 4  (2³² in other words 4.294.967.296 available addresses). Then, why to change?, this is an easy one, there are no more IPv4 address available. The last assigned block was assigned last February 2011. Then, if we want to continue communicating using IP (which by the way… IS internet) we must adopt this new version, having billions of possible addresses  (2¹²⁸ approximately 3.4×10³⁸).

IP v6 addressing

Same as in version 4, the basic configuration starts with a single IP address, a subnet mask and a default gateway. But beside many other “benefits” from this version, this 3 elements are integrated within the address itself; in other words, the IPv6 address contains this 3 basic elements on the numbering. This will render into a much simple solution with better routing and performance.

IPv6 is hexadecimal, that is we can use numbers from 0 to 9, plus the letters A to F (hexadecimal, or simply hex, is a base 16, or 16 bits, having A=10, B=11, C=12 and so on until F=15). An IPv6 address has 8 groups of 16 hex numbers, separated by a colon “:” having the first address as 0000:0000:0000:0000:0000:0000:0000:0000 and ending at ffff: ffff: ffff: ffff: ffff: ffff: ffff: ffff

Two examples:

1		2		3		4		5		6		7		8
2001	:	0db8	:	af01	:	b09c	:	0000	:	0000	:	0000	:	027d
fd04	:	eef3	:	0000	:	0001	:	0000	:	0000	:	a00f	:	0001

These addresses can be simplified. Any block containing zeroes (0000) can be expressed with a single zero (0), as we can see on the second example below, or it can even be more simplified by just having a single zero (0) representing the 3 contiguos blocks. And even more by just having the colons “:” (assuming all omitted blocks are zeroes

1

2

3

4

5

6

7

8

2001

:

0db8

:

af01

:

b09c

:

0000

:

0000

:

0000

:

027d

2001

:

db8

:

af01

:

b09c

:

0

:

0

:

0

:

27d

2001

:

db8

:

af01

:

b09c

:

:

027d

Following the example, the very same address can be expressed as:
2001:0db8:af01:b09c:0000:0000:0000:027d
2001:db8:af01:b09c:0:0:0:027d
2001:db8:af01:b09c::027d

Does Private address exist like in IPv4?

Yes, but because of the wide architecture of IPv6, it will not be necessary to use them (or at least meanwhile we live). There are mainly two of them, but there are reserved ranges and exceptions. All the rest of them are called Global Unicast Addresses or GUA.

LA (Local Addresses)

This is the equivalent to the auto-configuration APIPA (Automatic Private IP Autoconfiguration) which on version 4 is 169.254.0.0/16. This range has 2 main particularities:

1. Is NOT routable. Each configured segment remains within the network, and cannot be routed over public networks and/or interact with other networks.
2. Although there are too may addresses, there is a remote possibility to find duplication.

This range includes the loopback address (meaning myself or 127.0.0.1) ::1/128
The designated range is Fe80::/10

ULA (Unique Local Addresses)

The difference from LA addresses, is that these addresses can be routed, meanwhile the transmit devices (routers, switches, etc.)  are properly configured to accept such range. Explain in it differently, these addresses are not routable through public networks as internet, but they could be on private networks. The range fc00::/7 is equivalent to IPv4 ranges 10, 172.128 y 192.168. On the second example of IPv6 addresses mentioned before, one of this range address is used. A good page which might help us on these addresses is: http://www.simpledns.com/private-ipv6.aspx

IPv6 Subneting

We already explained that an IPv6 address has 3 parts.

1

2

3

4

5

6

7

8

Network Address

:

:

:

:

:

:

:

:

Subnet

:

:

:

:

:

:

:

:

Device Unique Address

2001

:

08db

:

af01

:

b09c

:

0000

:

0000

:

0000

:

027d

2001:db8:af01:b09c::27d

Network Address

The first 48 bits of an address, or expressed different, the first 3 blocks of an address (each block is a 16 bits or 4 hex characters). As a rule of thumb, each regional ISP will get a network address, which will be further sub-divided and assigned to other ISP and/or customers.

Subnet Address

The following 16 bits, or the 4th hex block. This single detail makes IPv6 way more effective than is ancestor, because communication wise, because the solely address already has the full routing information, source and target, avoiding routing calculations or having to modify the information once sent.

Unique Device Address

The last 64 bits of the address, or the last 4 blocks. This is the unique device identifier. Some devices can use the MAC address, but a DHCP can dynamically assign it regarding this number.