8845A/8846A
Programmers Manual
90
For any protocol, the packet consists of a destination address, a source address, some
protocol type information, a message length field, and error detection information.
The TCP/IP protocol adds additional fields to the packet for message sequence
determination and application port ID. The sequence number allows multi-packet
messages to be assembled into the correct order. The port ID specifies which application
running on a system is sending or receiving the message. Support of port IDs allow multi-
tasking within a computer system, with messages being sent and received from the
individual task.
In the TCP/IP protocol, error detection is supported by adding a checksum of the data in
the packet to the end of each packet. Using this checksum, the receiving system can
detect a bad packet and request a re-transmit (using the source address).
Each host (device attached to the network wiring) in an Ethernet network has a unique
address that is assigned by the interface hardware vendor. Each vendor is assigned a
block of addresses and it is required that each device produced have a unique Ethernet
address. Users of the TCP/IP protocol very rarely need to know or deal with the Ethernet
address.
When TCP/IP is used, each host is also assigned a unique IP address, but this address is
assigned when the TCP/IP protocol software is initialized. In contrast to the Ethernet
address, IP addresses have a structure that can be used by Network Management to
facilitate breaking the network into segments (subnets) that make better use of the
available data communication capacity (bandwidth). Use of IP addressing and the
network segmentation scheme is described in “IP Addresses and Segmented Networks”
in this appendix.
The relationship between a host’s Ethernet address and its IP address is established by the
TCP/IP protocol software. No user interaction is needed to establish this relationship.
However, some network management systems require that the Ethernet address of each
network host be kept in a database. So, it is sometimes necessary to communicate the
Ethernet addresses of all devices on the network to the Network Administrator.
IP Addresses and Segmented Networks
If all networks consisted of a few computers and other devices attached to each other
locally, use of IP addressing and network segmentation schemes would not be necessary.
However, networks have evolved past simple work groups into campus-wide, inter-
campus, and even international interconnect schemes. This makes it necessary to divide
the network into segments know as subnets.
The IP address provides two pieces of information: the network ID and the host ID. The
network ID must be unique among all network subnets within a connected Internet. It
identifies the network subnet to which a host is connected. The host ID must be unique
among hosts within a network subnet and identifies the host within the subnet.
Network subnets connected to the public Internet must obtain an official network ID from
a registrar approved by the Internet Network Information Center
(http://www.internic.net/) to preserve the unique network IDs within the Internet.
However, if you never intend to connect to the public Internet, you’re free to select your
own network IDs. The IP address is a 32-bit numeric address written in four numbers
separated by periods. Each of these four numbers can have a value in the range 1 to 254.
An IP address is written in the form w.x.y.z. For example, "198.178.246.10" (or
"C6.B2.F6.0A" in hexadecimal) is a valid IP address.
Table 35 lists the three commonly used classes of IP address assignments.
To Next Page
Loading...
