)
FIG. 23.1 OSI and TCP/IP compared.
| Previous | Table of Contents | Next |
The TCP/IP protocol stack represents a network architecture that’s similar to the ISO OSI networking model. Figure 23.1 shows the mapping of TCP/IP layers onto the ISO protocol stack.
FIG. 23.1 OSI and TCP/IP compared.
TCP/IP doesn’t make as fine distinctions between the top layers of the protocol stack as does OSI. The top three OSI layers are roughly equivalent to the Internet process protocols. Some examples of process protocols are Telnet, FTP, SMTP, NFS, SNMP, and DNS.
The Transport layer of the OSI model is responsible for reliable data delivery. In the Internet protocol stack, this corresponds to the host-to-host protocols. Examples of these are TCP and UDP. TCP is used to translate variable-length messages from upper-layer protocols and provides the necessary acknowledgment and connection-oriented flow control between remote systems.
UDP is similar to TCP, except that it’s not connection-oriented and doesn’t acknowledge data receipt. UDP only receives messages and passes them along to the upper-level protocols. Because UDP doesn’t have any of the overhead related to TCP, it provides a much more efficient interface for such actions as remote disk services.
The Internet Protocol (IP) is responsible for connectionless communications between systems. It maps onto the OSI model as part of the Network layer, which is responsible for moving information around the network. This communication is accomplished by examining the Network layer address, which determines the systems and the path to send the message.
IP provides the same functionality as the Network layer and helps get the messages between systems, but it doesn’t guarantee the delivery of these messages. IP may also fragment the messages into chunks and then reassemble them at the destination. Each fragment may take a different network path between systems. If the fragments arrive out of order, IP reassembles the packets into the correct sequence at the destination.
The Internet Protocol requires that an address be assigned to every device on the network. This address, known as the IP address, is organized as a series of four octets. These octets each define a unique address, with part of the address representing a network (and optionally a subnetwork) and another part representing a particular node on the network.
Several addresses have special meanings on the Internet:
CAUTION: It’s important that when you assign node numbers to your workstations, you don’t use 0, 127, or 255, because these are reserved numbers and have special meanings.
The IP addresses are assigned in ranges referred to as classes, depending on the application and the size of an organization. The three most common classes are A, B, and C. These three classes represent the number of locally assignable bits available for the local network. Table 23.4 shows the relationships among the different address classes, the available number of nodes, and the initial address settings.
| Class | Available Nodes | Initial Bits | Starting Address |
|---|---|---|---|
| A | 224=167,772 | 0xxx | 0–127 |
| B | 216=65,536 | 10xx | 128–191 |
| C | 28=256 | 110x | 192–223 |
| D | 1110 | 224–239 | |
| E | 1111 | 240–255 | |
Class A addresses are used for very large networks or collections of related networks. Class B addresses are used for large networks having more than 256 nodes (but fewer than 65,536 nodes). Class C addresses are used by most organizations. It’s a better idea for an organization to get several class C addresses because the number of class B addresses is limited. Class D is reserved for multicast messages on the network, and class E is reserved for experimentation and development.
Obtaining IP Addresses
The administration of Internet addresses is currently handled by the Network Information Center (NIC):
ON THE WEB: You also can reach the InterNIC on the Web at http://www.internic.net
When you connect a computer or a network to the Internet, in most cases your Internet service provider will be able to arrange for your network IP address registration.
Obtaining RFCs
In addition to assigning addresses, the NIC can provide other information of value. It’s a repository for all technical documentation related to the Internet. It has a collection of documents that describe all the associated protocols, routing methodologies, network management guidelines, and methods for using different networking technologies.
As mentioned in Table 23.1, RFC stands for Request For Comments. You can obtain RFCs from the Internet by using the FTP protocol to connect to several different repositories. The RFC series is available on the Internet via anonymous FTP from various sites, such as ftp.internic.net in the /rfc directory, and can also be accessed via Telnet at rs.internic.net.
| Previous | Table of Contents | Next |
){kind=link}