Detailed explanation of the differences between IPv6 and IPv4!

IPv6 is the abbreviation of Internet Protocol Version 6, where Internet Protocol is translated as "Internet Protocol". IPv6 is the next-generation IP protocol designed by IETF (Internet Engineering Task Force) to replace the current version of IP protocol (IPv4), claiming that it can encode a URL for every grain of sand in the world.

1. Basic Concepts of IPV6

As we all know, IPv6 uses a 128-bit address length and has a larger address space. First, let's take a look at what IPv6 looks like.

IPv6 datagrams are very different from IPv4:

• The type field of the data link layer (L2) is identified as 0x86dd, indicating that the upper layer protocol carried is IPv6 (IPv4 comparison: the type field is 0x0800);
• The header fields of IPv6 are very different from those of IPv4 (as you can guess, IPv6 and IPv4 are not compatible).

[[240144]]

Let's take a look at the IPv6 message header format:

IPv6 message header

The IPv6 header is simpler, with fewer fields. Compared with IPv4, there are several things to note:

• The IPv6 packet header is fixed length (fixed to 40 bytes), while the IPv4 packet header is variable length. This means that the efficiency of writing code to process IPv6 data packets will be greatly improved :);
• The meaning of the Hop Limit field in IPv6 is similar to the TTL in IPv4;
• The meaning of the Traffic Class field in IPv6 is similar to TOS (Type Of Service) in IPv4;
• The IPv6 packet header has removed the checksum field: removing this field is also an improvement on the IPv4 protocol. When an IPv4 packet is transmitted between networks, each time it is forwarded by a router, the TTL field is modified and the checksum needs to be recalculated. Since the checksums of the data link layer L2 and the transport layer L4 are already strong enough, removing this field in IPv6 will improve the forwarding efficiency of the router. It is worth mentioning that under the IPv6 protocol, the transport layer L4 protocols UDP and TCP are required to perform checksums (IPv4 is optional);
• The Next Header field in the IPv6 packet header indicates the "protocol type of the previous layer" or "extension header type".

The meaning here is very different from IPv4 and needs to be explained:

• When the IPv6 datagram carries the upper layer protocol ICMPv6, TCP, UDP, etc., the value of Next Header is 58, 6, and 17 respectively, which is very similar to the Protocol field in the IPv4 packet header;
• When it is not one of the above three protocol types, the IPv6 packet header is followed by an extension header. The extension header is a new concept introduced by IPv6. Each IPv6 data packet can carry 0 or more extension headers, which are organized in the form of a linked list. When an IPv6 data packet carries an extension header, the value of Next Header is the type value of the extension header.

Why do we need to introduce the concept of extension header? This is also one of the improvements of IPv6 over IPv4. The extension header replaces the optional information of IPv4, simplifies the header of IPv6, and enhances the extensibility of IPv6. Some students may wonder how to handle IPv6 fragmented data packets? In fact, it is the use of IPv6 extension header.

When sending a fragmented IPv6 datagram, IPv6 uses an extended header to organize the information of each fragment. The Next Header field value of the IPv6 message header is 44, indicating the presence of an extended header, which is the IPv6 fragment data information.

Compared with IPv4, fragmentation information is recorded in the fragment field of the IPv4 message header.

There are many types of IPv6 extension headers. In addition to the above-mentioned fragmentation header, there are also routing headers, hop-by-hop optional headers, etc. For details, please refer to RFC2460.

This chapter mainly introduces some very intuitive understanding of IPv6. The following will gradually introduce the basic knowledge and concepts of IPv6.

2. IPv6 address syntax

An IPv6 address uses colon hexadecimal notation: the 128-bit address is divided into 16-bit segments, and each 16-bit segment is represented in hexadecimal and separated by colons, for example:

A common public IPv6 address:

 2001:0D12:0000:0000:02AA:0987:FE29:9871

IPv6 addresses support the representation of compressed leading zeros. For example, the address above can be compressed as follows:

 200112:0:0:2AA:987:FE29:9871

To further simplify the IPv6 address, when there are several consecutive segments of 0 in the colon hexadecimal format, these segments can be compressed into a double colon representation. For example, the address above can be further simplified as follows:

 [pquote]200112::2AA:987:FE29:9871

For example, the IPv6 address FF80:0:0:0:FF:3BA:891:67C2 can be further simplified as follows:

 FE80::FF:3BA:891:67C2

It is worth noting here that the double colon can only appear once.

3. IPv6 address segmentation and prefix representation

IPv6 has a huge 128-bit address space. For such a large space, it is not divided arbitrarily, but divided into number segments according to bits (similar to the zone division algorithm of some 64-bit uin transformation and release numbers within Tencent).

The IPv6 address structure is as follows:

For example, RFC4291 defines n=48, m=16, which means that the subnet and interface ID each occupy 64 bits.

IPv6 supports a subnet prefix identification method, which is similar to the Classless Inter-Domain Routing (CIDR) mechanism of IPv4 (Note: IPv6 does not have the concept of a subnet mask).

Use the "IPv6 address/prefix length" representation method, for example:

• 2001:C3:0:2C6A::/64 indicates a subnet;
• And 2001:C3:0:2C6A:C9B4:FF12:48BC:1A22/64 represents a node address in this subnet.

As you can see, an IPv6 address consists of a subnet prefix + interface ID. The subnet prefix is ​​defined and allocated by the address allocation and management agency, while the interface ID can be generated by each operating system. The generation algorithm will be introduced in the following chapters.

4. IPv6 address types

There are three types of IPv6 addresses:

• Unicast, corresponding to ordinary public and private IPv4 addresses;
• Multicast, corresponding to the IPv4 multicast address;
• Anycast is a new address concept type added in IPv6.

IPv6 does not have a broadcast address, and uses a multicast address to implement the broadcast function. In fact, we are most likely to come into contact with unicast addresses in our work and life. The following article will focus on the types of unicast addresses. Students who are interested in multicast and anycast addresses can refer to the relevant RFCs and documents.