2nd Internet address structure [TCP/IP Detailed Volume 1: protocol]

This chapter describes the network layer addresses used in Interent, also known as IP addresses. Each device connected to the Interent has at least one IP address. The IP address is also required for devices that are used in a private network based on the TCP/IP protocol. In any case, the IP router implements a redirector that uses an IP address to identify traffic, and the IP address also represents the source of traffic.

Groups of IP addresses are assigned to users and organizations, who then assign them to devices, which are usually based on some "numbering schemes". For global interent addresses, a hierarchical structure management entity helps users and service providers assign addresses. An individual user is typically assigned an address by a interent service provider (ISP) to obtain an address and perform a route by paying a fee.

The most popular address type: IPV4 address. These addresses are usually made up of so-called dotted four or dotted decimal notation, such as 165.195.10.107. In IPv6, the length of the address is 128 bits, which is 4 times times the length of the IPv4 address. The traditional representation of a IPV6 address is to use a method calledblock or fieldThe four hexadecimal numbers, which are calledblock or fieldThe number of the numbers is delimited by colons.                                                                                                                         For example, a IPV6 address that contains 8 blocks can be written as 5f05:2000:80ad:5800:0058:0800:2023:1d71.   In some cases (for example, when you represent a URL that contains an address), the colon delimiter in the IPV6 address may be confused with other delimiters, such as the colon used between the IP address and the port number. In this case, enclose the IP address with the parentheses character [and].                                                                           TCP (Transmission Control Protocol)For example, url http://[2001:0db8:85a3:08d3:1319:8a2e:0370:7344]:443/refers to IPV6 host 2001:0db8:85a3:08d3:1319:8a2e:0 Port number 443 in 370:7,344 uses the HTTP, TCP, and IPV6 protocols.

Because of the large number of addresses (especially IPv6), it is convenient to divide the address space into blocks. IP addresses can be grouped according to type and with small. Most IPV4 address blocks are eventually subdivided into an address that identifies a computer network interface that connects Interent or some dedicated intranet. These addresses are called unicast addresses , and other types of addresses include broadcast, multicast, and anycast addresses, which may involve multiple interfaces and some special-purpose addresses.

When the Interent address structure is initially defined, each unicast IP address has a network part that identifies the IP address used by the interface to be discovered in which network, and a host address , Used to identify specific hosts in the network that are given by the network section. Therefore, some contiguous bits in the address are called network numbers , and the remaining bits are called host numbers . At that time, most hosts had only one network interface, so the term interface address and host address were sometimes used interchangeably.

In reality, different networks may have different numbers of hosts, and each host needs a unique IP address. A partitioning method is to assign different sizes of IP address space to different sites based on the current or expected number of hosts. The division of address space involves five main categories . Each class is based on the number of hosts that can be accommodated in the network, determining the number of bits assigned to the network number and host number in a 32-bit IPV4 address.

The IPV4 address space is initially divided into five main categories. Classes A, B, and C are used to assign addresses to interfaces in Interent (unicast addresses), and in some other special cases. Classes are defined by the first few in the address: 0 is Class A, 10 is Class B, 110 is Class C, and so on. Class D (1110) addresses are for multicast use, and Class E (1111) addresses are reserved.

Category Address structure:

Class A: 0.0.0.0~127.255.255.255

Class B: 128.0.0.0~191.255.255.255

Class C: 192.0.0.0~223.255.255.255

Class D: 224.0.0.0~239.255.255.255

Class E: 240.0.0.0~255.255.255.255

Assigns a network number to a site after it is connected to the Internet, and then further divides the local subnet number by the site administrator. Implementing this idea requires changing the limits of the network and host parts of an IP address, but doing so is only for a site itself; the remainder of Interent will only "see" The traditional Class A, class B, and C sections, and the methods that support this feature are called Subnet addressing . With subnet addressing, a site is assigned a network number of Class A, Class B, or Class C, leaving some remaining host numbers for further intra-site allocations. The site may further divide the host portion of the base address into a subnet number and a host number. In essence, subnet addressing adds an extra portion to the IP address structure, but it does not add a length to the address.

Subnet addressing provides additional flexibility at the expense of increased costs. Because the definition of the current subnet field and host field is specified by the site (not determined by the network number classification), all routers and hosts in a site need a new way to determine the subnet portion of the address and the host part of it. Before the subnet, this information can be obtained directly from a network number, just know is a class, Class B or Class C address (by the address of the first few representations).

A subnet mask is an allocation bit used by a host or router to determine how to obtain network and subnet information from a host's corresponding IP address. The IP subnet mask is the same length as the corresponding IP address (IPV4 is 32 bits and IPV6 is 128 bits). They are typically configured in the same way as IP addresses in a host or router, either statically (usually a router) or in a dynamic manner, such as Dynamic Host Configuration Protocol (DHCP). The subnet mask can be expressed in dotted decimal notation (255.255.254.0) or as a prefix length (/23).

Masks are used by routers and hosts to determine the end of a network/subnet portion of an IP address and the start of the host section.

Note that the rest of the Interent routing system does not require the knowledge of the subnet mask, because routers outside the site make routing decisions based only on the network number portion of the address, and do not require a network/subnet or host part. Therefore, the subnet mask is purely a local problem within the site.

We found that different lengths of subnet masks can be applied to the same network number in different parts of the same site. While this increases the complexity of address configuration management, it also improves the flexibility of the subnet structure, because different subnets can accommodate a different number of hosts. Currently, most hosts, routers, and routing protocols support variable-length subnet masks (VLSM).

Note that each interface of the host and router needs to be described with an IP address and a subnet mask, but the mask determines the difference in the network topology. Based on the dynamic routing protocols running in the router (for example, OSPF, Is-is, RIPV2), traffic can flow correctly between hosts in the same site, and from the Internet to or from external sites.

In each IPv4 subnet, a special address is reserved as the subnet broadcast address. The subnet broadcast address is formed by setting the network/subnet portion of the IPV4 address to the appropriate value, and all bits of the host part set to 1. The subnet broadcast address is constructed by inverting the subnet mask (changing all 0 bits to 1 and vice versa) and bitwise OR operations with the address (or equivalent network/subnet prefix) of any computer in the subnet.

Historically, the use of broadcast addresses as destination datagrams is also known as directed broadcasts . At least theoretically, this broadcast can be routed over the internet as a single datagram until it reaches the destination subnet, and is sent to all hosts in the subnet as a set of broadcast datagrams.

In addition to the subnet broadcast address, the special purpose address 255.255.255.255 is reserved for local webcasts (also known as Limited broadcasts ), which are not forwarded by routers at all.

Broadcast addresses are typically used with certain protocols, such as UDP/IP or ICMP, because these protocols do not involve a two-way conversation with TCP/IP. The IPV6 does not have any broadcast address, the broadcast address can be used in IPv4, and IPV6 uses only the multicast address.

The IPV6 address uses a special prefix to represent an address range . A IPV6 address range is the size of the network it is available in. Important examples of scope include node-local (for communication on the same computer only), link-local (only for nodes in the same network link or IPv6 prefix), or global (Internet-wide). In IPv6, most nodes typically use multiple addresses on the same network interface.

The link-local IPV6 address (and some global IPV6 addresses) uses the interface identifier (IID) as the basis for the allocation of a unicast IPV6 address.

In the IEEE Standard, EUI represents the expansion of a unique identifier . The EUI-64 identifier begins with a 24-bit organization Unique identifier (OUI), followed by an organization-assigned 40-bit extension identifier , which is identified by the preceding 24 bits.

To help alleviate the stress of IPV4 addresses (especially Class B addresses), categorical addressing schemes typically use a vlsm-like scheme to extend the Internet routing system to support classless Inter-Domain Routing (CIDR).

For IPv4 and IPV6, no address ranges specified as special, multicast, or reserved addresses are available for unicast use. Some unicast address spaces are reserved for building private networks. Addresses from these scopes can be used for communication between hosts and routers within a site or organization, but not across the global Internet. Therefore, these addresses are sometimes referred to as non-routable addresses. In other words, they cannot be routed in the public Internet.

IPV4 and IPV6 support multicast addressing. An IP multicast address (also known as a group or group address) identifies a set of host interfaces, not a single interface. In general, a group can span the entire Internet. The network part covered by a group is called the scope of the group. Common scopes include node-local (same computer), link-local (same subnet), site- local (for some sites), global (entire Internet), and management . The address of the managed scope can be used for addresses that have been manually configured to the router within a network region. Site administrators can configure routers to manage scope boundaries , which means that multicast traffic for related groups is not forwarded by the router. Note that site-local and administrative scopes are only valid when using multicast addressing.

Under the control of the software, the protocol stack in each Internet host can join or leave a multicast group. When a host sends data to a group, it creates a datagram that uses the (unicast) IP address as the source address, using the multicast IP address as the destination address. All hosts that have joined the group will accept any datagrams sent to that group.

At this point, the original multicast service model has become known as any source multicast (ASM). In this model, any sender can be sent to any group, and the recipient of a join group is assigned a unique group address. A new scheme is called Source-specific multicast (SSM), and only one sender is used in each group. In this case, when a host joins a group, it is assigned a channel address, which includes a group address and a source IP address. SSM avoids the complexity of ASM model deployment. Although a variety of multicast forms are widely used throughout the Internet, SSM is currently a more popular candidate.

The anycast address is a unicast IPv4 or IPV6 address that identifies different hosts depending on the network on which it resides. This is accomplished by configuring the router to notify multiple sites in the Internet that there is the same unicast route. Therefore, an anycast address does not refer to a host in the Internet, but rather to a host that is "most suitable" or "closest" to the anycast address. Anycast addresses are most commonly found in a computer that provides a common service.

The IP address space is usually allocated as a large block, which is done by a number of hierarchical organizations of authoritative bodies .

For unicast IPv4 and IPV6 address spaces, the IANA assigns permissions primarily to several regional Internet registries (RIR).

In IPv4 and IPv6, multicast addresses (that is, group addresses) can be described according to their scope, and they need to determine the multicast mode (static, dynamic protocol or algorithm), and whether ASM or SSM is used.

"Summary":

An IP address is used to identify and locate the network interface of a device in an entire Internet system (unicast address). It is also used to identify multiple interfaces (multicast, broadcast, or anycast addresses). Each interface has a minimum of 32-bit IPV4 addresses, and there are typically several 128-bit IPV6 addresses.

Both IPV4 and IPV6 support address formats that point to multiple network interfaces at the same time. IPV4 supports broadcast addresses and multicast addresses, but IPV6 only supports multicast addresses. Broadcasts allow one person to communicate with everyone, while multicast allows one person to communicate with multiple people.

It can be said that the development and deployment of CIDR ( classless Inter-Domain routing ) is a fundamental change in the Internet core routing system. CIDR successfully provides more flexibility in allocating the address space and provides the scalability of the route through aggregation.

2nd Internet address structure [TCP/IP Detailed Volume 1: protocol]