introduction
With the rapid growth of the Internet and the gradual depletion of IPv4 address space, IPv6 as the next version of the Internet protocol, the ultimate replacement of IPv4 will inevitably become inevitable. In this series, we will discuss the problems existing in the current Internet and the IPv6 resolution, IPv6 address assignment, new IPv6 Baotou and its extensions, IPv6 replaces ICMP protocol and IGMP protocol, neighboring node interaction, and IPv6 address automatic configuration Wait.
IPv6 overview
The current IPv4 has not changed since the RFC 791 standard in 1981. It turns out that IPv4 has a relatively strong vitality, easy to achieve, and has good interoperability, withdrawn from early small-scale internet to expand to today's global Internet applications. All of this should be attributed to IPv4 initial excellent design.
However, there are still some development that has not been expected at the beginning of design:
In recent years, Internet has expired rapid development in an exponential level, causing the IPv4 address space to be depleted. The IP address becomes more rare, forcing many companies have to use NAT to map multiple internal addresses into a public IP address. Although the address conversion technology alleviates the lack of public IP address to some extent, it does not support certain network-level security protocols, and it is inevitable that there are any new problems in address mapping, which has caused some new problems. Moreover, relying on NAT and impossible to fundamentally solve the problem of lack of IP address, with the sharp increase in networking equipment, IPv4 public addresses will always be completely exhausted.
The Internet main network router maintains an enhancement of large routing capabilities. The current IPv4 routing basic structure is a mixture of plane routing mechanisms and hierarchical routing mechanisms, and the Internet core main network router can maintain a routing entry of more than 85,000.
Address configuration tends to be more simplified. At present, most IPv4 address configurations require manual operation or using DHCP (Dynamic Host Configuration Protocol) address configuration protocols. As more and more computer and related devices use IP addresses, it is inevitable to improve the degree of automation of address configurations, making it more simplistic, and other configuration settings do not depend on the management of DHCP protocols.
The growth of IP layer security needs. Special data communication in public media such as Internet generally requires encryption services to ensure that data will not be disclosed or stolen during transmission. Although there is currently an IPSec protocol to provide security protection for IPv4 packets, because the agreement is only an optional standard, the company uses the individual private security solutions to usually quite common.
Better real-time QoS support needs. The IPv4 QoS standard is dependent on the IPv4 service type field (TOS) and authentication using UDP or TCP ports on real-time transmission support. However, IPv4's TOS field function is limited, and it may cause real-time transmission timeout factors and too much. In addition, if IPv4 packets are encrypted, the TCP / UDP port cannot be authenticated.
In order to solve the above problem, the Internet Engineering Task Group (IETF) has developed IPv6. This new version has also been known as the next generation IP, combines multiple proposals for IPv4 upgrades. In design, IPv6 dynamism avoids adding too much new features to reduce the impact on existing high-level and low-level protocols as much as possible.
IPv6 characteristics
Some new features of IPv6 protocols are listed below:
New Baotou Format
Larger address space
Efficient level addressing and routing structure
Full state and stateless address configuration
Built-in security facilities
Better QoS support
New protocol used for neighboring node interaction
Scalability
These new features are discussed in turn:
New Baotou Format
The design principle of the new IPv6 Baotou is to try to minimize the header overhead. The specific practice is to remove some non-critical fields and optional fields out of the header. In the extended clasp header after IPv6 Baotou, even though IPv6 address length is IPv4 four Femide, but the header is only twice the IPv4. Improved IPv6 Baotou handles efficient in the centering router.
Since both the header is not interoperability, and IPv6 is not a functional extension set of IPv4, IPv4 and IPv6 must be implemented in the host and router to identify and process these two cladding formats. Larger address space
IPv6 address length is 128 (16 bytes), that is, 2 ^ 128-1 (3.4e 38) addresses, this address space is 1E28 times the IPv4 address space (or in the current global total number of people, per capita Assign 1.8 × 1019 IPv6 addresses). IPv6 uses grading address mode to support multi-stage subnet address allocation from ITERNET core to enterprise internal subnets.
In the huge address space of IPv6, the current global networking device has accounted for only a small part, and there is sufficient margin for future development. At the same time, due to sufficient address space, address conversion technology such as NAT will no longer need.
Efficient level addressing and routing structure
IPv6 uses clustering mechanisms to define a very flexible hierarchical addressing and routing structure. The multiple networks at the same level are represented in the upper router as a unified network prefix, which can significantly reduce the routing entry that the router must maintain. Ideally, a core trunk router is only maintained without more than 8192 entry. This greatly reduces the router's search and storage overhead.
Full state and stateless address configuration
To simplify host configuration, IPv6 supports the full state and stateful and stateless. In IPv4, the Dynamic Host Configuration Protocol DHCP implements automatic settings for host IP addresses and their related configurations, IPv6 Repair IPv4's automatic configuration service, referred to as full state autoconfiguration. In addition to the full state automatic configuration, IPv6 also uses an automatic configuration service called stateless autoconfiguration. During the stateless automatic configuration, the online host automatically obtains the address prefix of the local router and the link local address and related configuration.
Built-in security facilities
IPv6 fully supports IPsec, which requires a standard network security solution to meet and improve the cooperative work capabilities between different IPv6 implementations.
Better QoS support
The new field of the IPv6 Baotou defines how the data stream recognizes and processes. The flow identifier in the IPv6 Baotou is used to identify the data stream identity, using this field, IPv6 allows the end user to request the communication quality. The router can identify all packets belonging to a particular data stream based on this field and provide specific processing on the need to provide specific processing. Since the data stream identity information is included in the IPv6 poppet, QoS support can also be obtained even if the packets encrypted by IPSec encrypted.
New protocol used for neighboring node interaction
IPv6 neighbor find protocol (Neighbor Discovery Protocol) uses a range of IPv6 Control Information Packets (ICMPv6) to implement interaction management of neighboring nodes (nodes on the same link). Neighbor Discovery Agreement and Efficient Multicast and Unicast Neighbor Discovery Replacement The ICMPv4 Router Discovery and ICMPv4 Router Discovery Packets are replaced by the inline broadcast address resolution protocol ARP.
Scalability
IPv6 features have strong scalability, and new features can be added to the extended header after IPv6 headers. Unlike IPv4, the Baotou can only support 40-byte options, the size of the IPv6 extension header is only limited by the maximum number of bytes of the entire IPv6 package.
IPv4 and IPv6 differences
Below we will list the main differences of IPv4 and IPv6 in Table 1 for comparison.
Table IPv4 and IPv6 main differences
IPv4
IPv6
32 digits
128 digits
IPsec is an optional expansion protocol
IPSec becomes an integral part of IPv6, supporting IPSec is a must-support data stream recognition item in the header
The flow identity field in the bag head provides a data stream identification function, supports different QoS requirements.
Both the router and send the host to complete the segmentation
The router no longer performs segmentation, and segmentation is only performed by the send host.
Baotou includes integrity checking and
Does not include integrity checking in the header
Baotou contains optional options
All optional content moved to the extended clasp
ARP protocol uses broadcast ARP request frames to parse IPv4 addresses
Multicast neighbor request message replaces ARP request frame
IGMP protocol is used to manage local subnet members
Alternative IGMP Management Local Subnet by MLD Packet
ICMP router discovers an optional protocol to determine the IPv4 address of the best default gateway
ICMPv6 router request and router release packet as required protocol
Use the broadcast address to send data to all nodes of subnet
IPv6 no longer has broadcast addresses, but uses multicast addresses for all nodes in the link partial range.
Address configuration mode for manual operation or through DHCP protocol
Address Auto Configuration
In the DNS server, the IPv4 host name is generated by the mapping of the address using a resource record type.
IPv6 host name and address mapping uses new AAAA resource record types to establish
IN-addr.arpa domain provides IPv4 address - hostname resolution service
IP6.Int domain provides IPv6 address - hostname resolution service
Support 576-byte packets (possibly segmentation)
Support 1280 byte packets (no segmentation)
IPv6 and LAN
In IPv6, the data link layer frame structure includes three parts: data link layer header and newsshooting, IPv6 Baotou, payload, see Figure 1.
Figure 1 IPv6 packet on the data link layer
Among the typical local area network technologies such as Ethernet, token ring and FDDi networks, IPv6 has two packages: Ethernet II packages or applications in IEEE 802.3, IEEE 802.5, and FDDI SNAP (Subnet Access Protocol) package.
Ethernet II package
When encapsulated with Ethernet II, the EtherType field value in the Ethernet II header is: 0x86dd (this field value = 0x800 represents IPv4 packet). Using the Ethernet II package, IPv6 package minimum 46 bytes, up to 1500 bytes. Figure 2 is an Ethernet II package for an IPv6 package.
Figure 2 Ethernet II package
IEEE 802.3, IEEE 802.5, and FDDI package
In IEEE 802.3 (Ethernet), IEEE 802.5 (Token Ring Network) and FDDI network, the EtherType field value is: 0x86DD. Figure 3 is a SNAP package for an IPv6 package.
Figure 3 SNAP package
Using SNAP packages in Ethernet, IPv6 package minimum 38 bytes, maximum 1492 bytes, in the FDDI network, the IPv6 package is up to 4352 bytes. For the maximum number of bytes of IPv6 packages in the token ring network, interested readers can refer to RFC 2470.
