Transitioning from IPv4 to IPv6, you can't miss these knowledge points

Preface

Network is one of the basic skills for engineers. After all, we can't do anything without it! From the recruitment requirements of major companies, network is also a key assessment part. Today, let's review (repay) the college courses (debts) together! IP protocol of TCP/IP protocol suite.

[[275868]]

Understanding the network model

The network designers came up with a layered design that separates each function and delivers it to different layers. This has the advantage of facilitating updates and maintenance (the advent of IPv6 did not require a redesign of the entire network), and also facilitates our learning and understanding. Below is a diagram of the structure of each version of the network system:

Let's clarify the responsibilities of each layer (taking the five-layer protocol as an example):

Application layer: This layer protocol is mainly responsible for the rules of interaction between various applications. For example, HTTP and DNS, the data unit of their interaction can be called "message".

The transport layer provides general data transmission services for the application layer. It runs errands for the upper layer. TCP and UDP are both transport layer protocols.

The network layer provides communication services for different hosts in the network. The IP protocol works at this layer.

The data link layer is responsible for data transmission between two adjacent points.

Physical layer, transmission bit stream, 0 or 1. Each layer encapsulates the data into a data format that it can understand and delivers it to the next layer. The next layer will receive the data as its own data and add the necessary data of this layer again, and deliver it downward until the link layer. After being delivered to the destination, each layer removes the necessary data of the layer and delivers it to the upper layer. As shown in the following figure:

Understanding IP protocol

After understanding the network model and the direct cooperation of each layer, let's move on to today's topic, Internet Protocol (IP).

This section mainly focuses on understanding what functions the IP protocol provides us, in other words, what it can do.

• Provide an ID card for the host (PC or router or gateway) in the network.
• Defines the basic communication method between network hosts, which can be seen from the format of IP datagram.
• It provides convenience for routing addressing, as can be seen from the routing table structure.

Any protocol can be composed of three parts: syntax, semantics, and synchronization.

Syntax, the format of information exchange, for IP protocol, is the format of IP message.

Semantics, that is, what kind of control information needs to be sent and how to respond after receiving the information.

Synchronization, a detailed description of the sequence of events.

That is, it defines the rules for communicating using the IP protocol.

IPv4

IPv4 is the 4th version of the IP protocol, which is what we use now. Let's learn about this version from multiple aspects.

IP address

All hosts on the network need an IP address so that they can know each other's location. IPv4 addresses are composed of 32 bits and are often expressed in dotted decimal notation (192.168.1.1).

Classification

The classification of IP addresses is to divide all IPs into several categories, each of which is composed of fields of fixed length. They are network number and host number. The following is a schematic diagram of the composition of various types of addresses:

• Class A address, 8-bit network number, the first bit is fixed to 0, there are only 126 available network numbers (2^7-2), network numbers from 0 to 127. Subtract 2 because: IP address all 0 means "this network", reserved; network number 127 is used for software loopback testing, if the host sends data with the destination 127.0.0.1, the data will not be sent to any host on the network. That is to say, addresses starting with 127 cannot be used. For each network number in the Class A address, there are (2^24-2) corresponding host numbers: host number all 0 means the network address corresponding to its network number, all 1 means all hosts.
• The network number of a Class B address is 16 bits, but the first two bits are fixed to 1 0. No matter how the value is taken, it cannot be all 0 or all 1. The 128.0.0.0 network is not used, and the smallest Class B network address actually used is 128.1.0.0. Therefore, there are (2^14-1) Class B network addresses.
• Class C addresses have 24-bit network numbers, the first 3 bits are fixed to 1. The network address of 192.0.0.0 is also not used. The smallest Class C address is 192.0.1.0.

To summarize:

Benefits of classification:

• Convenient management, the IP address management agency only needs to manage the network number, and the rest can be managed by subordinates.
• When routing, packets are forwarded according to the network address to reduce the routing table.

Datagram Format

First look at the basic format of IP datagram:

You can see that the IP datagram contains a header and a data part. The header contains a fixed 20 bytes and a variable part. The following is an explanation of each part:

• Version, occupies 4 bits, the version number of the IP protocol, both communicating parties need to be the same.
• The header length occupies 4 bits, and the unit is 4 bytes. The maximum that can be represented is (2^4-1) * 4 = 60 bytes. When the header length of the IP packet is not an integer multiple of 4, it needs to be filled with a padding field.
• Differentiated services, occupies 8 bits, indicates the service type, not used.
• Total length, 16 bits, represents the total length of the header and data (in bytes). Therefore, the maximum total length of a datagram is (2^16-1 = 65535) bytes.
• Identification, 16 bits. When the length of a datagram exceeds the MTU of the lower layer, it needs to be fragmented. The fragmented datagrams have the same identification so that they can be reassembled.
• Flag, occupies 3 bits, only 2 bits are meaningful at present
• The lowest bit, MF (More Fragment) = 1 means there are more fragments to come. MF = 0 means the last one in the fragment.
• The middle one, DF (Don't Fragment), does not fragment, and fragmentation is allowed only when DF=0
• Fragment offset, 13 bits. The relative position of a fragment in the original packet after the longer packet is fragmented. Must be an integer multiple of 8 bytes.
• Survival time, occupies 8 bits, Time To Live (TTL), initially refers to the number of seconds of survival, and later becomes the number of hops through the route. Every time it passes through a route, TTL decreases by 1. When it is 0, the data is discarded. Obviously, the maximum number of hops is 255. When it is 1, it can only be transmitted in the local area network.
• Protocol, occupies 8 bits. It indicates which protocol is used to transmit the data in the datagram, so that the IP layer of the destination host knows which protocol to deliver the data to. Commonly used protocols and corresponding values:

• Header checksum, occupies 16 bits. The header checksum of the datagram does not include the data part.
• Source address, destination address, each occupies 32 bits

IPV6

The emergence of IPv6 is to solve the problem of IPv4 resource exhaustion. It still supports connectionless transmission, but transmits PDU packets instead of IPv4 datagrams.

IP address

IPv6 addresses are 128 bits, which is enough for the foreseeable future. However, they are not compatible with IPv4. If host A using IPv4 and host B using IPv6 communicate, special processing is required. The method will be introduced later.

IPv6 Address Representation

IPv6 addresses are written in colon hexadecimal notation, with 8 groups of 4 hexadecimal digits separated by colons. For example:

6845:8C64:FFFF:FFFF:0:1180:9000:FFFF

• If all the numbers in a group are 0, three of them can be omitted.
• If there are multiple consecutive groups of 0, you can use a colon to replace them, but you can only compress them once. FF05:0:0:0:0:0:0:B3 can be compressed to FF05::B3 1090:0:0:0:8:800:200C:417A can be compressed to 1090::8:800:200C:417A All zeros can be recorded as::
• The last two groups can be represented using IPv4 dotted decimal notation 0:0:0:0:0:0:128.10.2.1 which can be represented as ::128.10.2.1

Datagram Format

IPv6 datagrams consist of two parts, the basic header and the payload. The payload runs with 0 or more extension headers.

As you can see, the IPv6 header is a fixed 40 bytes, which is different from IPv4; there are also significant differences in composition. Let's take a look at the meaning of each part.

• Version, occupies 4 bits, the version of the protocol, both parties need to be the same
• Traffic class, occupies 8 bits. Distinguishes different IPv6 datagrams or priorities.
• Flow label, occupies 20 bits. A new mechanism of IPv6 is to support resource pre-allocation, running routing to allocate and associate each datagram with specific resources. IPv6 proposes the concept of flow. It refers to a series of datagrams from a specific source point to a specific destination on the Internet (such as real-time audio or video transmission). Routers on the path of this flow guarantee the specified quality of service. All datagrams belonging to the same flow have the same flow label. Assigning a value of 0 means closing.
• Payload length, 16 bits. Indicates the number of bytes excluding the basic header. Maximum 64KB.
• The next header, occupies 8 bits. It is equivalent to the protocol field or option field of IPv4.
• When IPv6 has no extension header, this field is the same as the protocol field of IPv4.
• When there is an extension header, it indicates the type of the first extension header.
• Hop limit, same as IPv4 TTL
• Source address, destination address, each occupies 128 bits

Transition from IPv4 to IPv6

The network is so large that the transition from IPv4 to IPv6 cannot be completed overnight. To communicate under the two versions of the protocol, there are two methods:

Dual protocol stack

Simply put, a host can understand two versions of content, so the host must also have two versions of the corresponding IP address. A host with a dual protocol stack can know the protocol version used by the destination host through the DNS system.

The following is a schematic diagram of two hosts using IPv6 communicating. They need to communicate through the IPv4 network, and when necessary, the dual-stack host performs protocol conversion. Of course, some information may be lost during the conversion, which is also inevitable.

Using Tunneling Technology

The principle of tunneling technology is that when IPv6 data needs to enter the IPv4 network, the IPv6 datagram (PDU to be precise) is treated as the data part of the IPv4 datagram, and transmitted using the IPv4 version. When leaving the IPv4 network, it is assembled into IPv6 data and sent to the destination. The following is a schematic diagram:

Conclusion

There is still a lot to talk about the IP protocol, so I won’t go into detail here. I hope you will have a clearer understanding after reading this.