Chapter 4, 5: Classification of Internet addresses and ARP-Inter-network interconnection with TCP/IP

refer to "Internet Interconnection with TCP/IP-Volume One", chapters 4th and fifth

Section 4 Chapter: Classification of Internet addresses

4.3 the initial classification addressing scheme

TCP/IP each host on the Internet is assigned a unique + bit bit Internet address, which is used in all communication with the host.

4.4 indicate network connection by address

because An IP address is an address to a network and to a host on that network, so one of the addresses does not indicate a computer, but instead indicates a connection to the computer to the network. Therefore, a router that is connected to n Networks has several different IP addresses, each of which corresponds to a network connection.

4.5 network address and directed broadcast address

by convention, the host number 0 is never assigned to a separate host, and the host number portion of the IP address of 0 is used to point to the network itself.

The Internet address can be used to point to a host, or to a network, according to the rules, the host number of all bits are 0 of the address is reserved to the network itself. The address that the host number consists of all 1 is reserved for directed broadcasts.

4.10 defects in Internet addressing

If a host is moved from one network to another, its The IP address must be changed.

because forwarding uses the The network portion of the IP address, so when the packet is sent to a host with multiple IP addresses, the path used by the packet transfer is related to the address used.

4.13 Summary of Special address conventions

4.18 Summary

TCP/IPUse +bit binary address as a generic machine identifier, calledNetwork Protocoladdress orIPaddress, the identifier can be divided into two parts: prefix identifiercountedconnected to the network, the rest of the network of the Grate machine provides aonlyidentifier. The initialIPthe addressing method is called a classification method, and each prefix is assigned to one of three basic classes. The leading bits define the class to which the address belongs, and the capacity of each type is different. The classification method provides127network that connects millions of hosts, provides thousands of network connections to thousands of hosts, and provides the maximum number of hosts254millions of networks. For ease of understanding, the address is written in dotted decimal notationaddress, put4athe value of the eight-bit group is written as a decimal number, with a smallnumberpoints apart.

ip The address encodes the network representation and the identifier of the specific host on the network, so the forwarding efficiency is high. An important property of addresses is that they refer to network connections. There are multiple addresses for hosts with multiple connections. An advantage of the Internet addressing approach is that this form packages a specific host, address of a network, or There is a host ( broadcast) address. ip The biggest drawback of addressing method is that if a machine has multiple addresses, When the path to a pending interface does not exist (for example, a specific network cannot use the Just know that one of the addresses of this machine may not be enough to reach it.

In order to be able to exchange binary data between machines , TCP/IP Protocol Requirements the integer in the field uses the a standard byte sort. Before the host is sent, all binary numbers must be swapped from the internal format to the network standard byte order, and then converted from the network standard byte order to the internal format after receiving the packet.

Section 5 map of Internet addresses to physical addresses ( ARP )

5.3 two types of physical addresses

Fixed long addresses (such as the address used by Ethernet) and easy-to-configure short addresses

5.5 parsing with dynamic bindings

through the Address Resolution Protocol, which is APP, the host can find the physical address of the destination station as long as it knows the IP address ofa destination host on the same physical network .

5.8 ARP the Improvement

each The ARP broadcast packet contains the binding of the sender's IP to the physical address , and the receiver is now updating the sender IP-to-Physical address mapping in its own cache before processing the ARP packet (for target processing) .

5.9 ARP relationships with other protocols

ARP is an underlying protocol that hides the underlying network address addresses and allows the allocation of arbitrary IP addresses to each machine, and we treat ARP as part of the underlying physical network system, rather than part of an Internet system.

5.11 ARP the encapsulation and identification

when an ARP message is sent from one machine to another machine, it must be placed in a physical frame .

5.12 ARP format

Hardware Type (HARDWARE TYPE) field indicates the type of hardware interface The sender wants to know, and for Ethernet, the field contains a value of1. Protocol type (Procotl TYPE) field indicates the high-level protocol address type provided by the sender;Ipaddress, which contains a value of0800 (16in-process). Operation (Opeartion) field indicates whether theArpRequest(1),ArpResponse(2),RARPRequest(3),RARPResponse(4). Hardware address Length (Hlen) field and protocol address length (Plen) field allowsArpapply on any network because they indicate the hardware address and the length of the High-level protocol address, respectively. If the sender knows their hardware address andIpaddress, the sender's hardware address (SENDER HA) and the sending partyIpAddress (SENDER IP) field gives the hardware address andIpaddress.

5.15 Summary

the assignment of an IP address is independent of the physical hardware address of the machine.

by Radio ARP request, a machine can use ARP to find the hardware address of another machine, and the request contains the IP address of the host where its hardware address needs to be found . All machines in the network are subjected to ARP requests, and if the IP address of a machine matches the request, the machine issues a response with the required hardware address, which is sent directly to a machine, not broadcast.

to improve ARP efficiency, each machine to the IP address to the physical address of the binding to save in the tell cache, because the Internet traffic tends to become a series of inter-machine interaction, so the cache excludes most ARP broadcast requests Early confirmation can eliminate the jitter phenomenon.

RARP is an old protocol related to ARP it allows the computer to obtain an IP address when the system starts .

Chapter 4, 5: Classification of Internet addresses and ARP-Inter-network interconnection with TCP/IP