What is Internet Protocol For?

Internet Protocol

The Internet Protocol (IP) is the basis upon which all other Internet protocols operate. IP provides the basic mechanism for the forwarding of data between two computers on separate networks. IP can also fragment packets it they are too large for some older networks to forward, but this feature is largely obsolete because all routers built in the past decade are able t pass large IP packets.

IP pickets are simply handed from computer to computer until they reach their destination. The computer sending the packet and the computer receiving

the packet are called end systems because they are at the ends of the communication on session. The computers between the end systems are called intermediate systems. Intermediate system is a generic and more theoretical term for computers more commonly called routers, gateways, or multihomed hosts.

Router usually refers to specially designed computers optimized for routing packets.

Gateway refers to general-purpose computers that are simply used as routers and perform no other function.

Multihomed host refers to general-purpose computers that perform some other function in addition to routing packets such as file service or Internet site hosting.

The term host is commonly used in the UNIX and Internet communities to

mean any computer (client, server, or peer) that is directly attached to the Internet.

Services of IP

IP provides the functions of addressing and fragmentation only to support

packet forwarding; it does not presume or implement any other functionality.

• IP cannot guarantee that a packet will reach its destination.

• IP has no ability to perform flow control.

• IP performs no error correction.

• IP performs no error detection for the data payload.

• IP does not guarantee that packets will arrive in order and does not order them sequentially.

IP relies on the data link to transmit data in an error-free condition and

does not attempt to provide any guarantees of service. Other protocols, which are transported withi
IP packets, add information such as packet serial numbers and error-correction codes. The destination system can check to see if all the packets have arrived, arrange them in the correct order, and request that

any missing packets be sent again based on this additional information. TCP performs all these functions, as explained in the next section.

IP treats each packet as an independent entity, unrelated to any other packet being transmitted. IP does not have logical or virtual connections, circuits, sockets, or any other mechanism to provide associations between packets. These functions are all provided by higher (or occasionally lower) level protocols.

IP does not perform error correction. IP does, however, implement limited error checking to verify that the header information is correct; damaged header information could result in the packet being forwarded to the wrong address. If, at any time, a router on the path between the sending computer and the receiving computer detects that an IP packet's address header has become damaged (by comparing the header with the header's checksum), the router will simply discard the packet without notification of any kind. Again, higher- level protocols will determine what data is missing and generate a request for retransmission. This header checksum does not detect errors that may have crept into the data portion of the packet; that function also is the responsibility of higher-level protocols.

IP does include information about how long a packet should remai
"alive" in a system. Every IP packet contains a time-to-live indicator and decreases this counter by one each time a router forwards the packet or whenever one second of real time elapses. The time-to-live counter usually starts at 64, 128, or 255. When a packet's time-to-live counter reaches zero, the packet is discarded.

This event can occur in three (rare) cases:

• When the network is too busy to forward packets in a timely manner

• When a circular route exists and packets are simply being passed around it

• When the route between two computers is simply too long to he useful

In all three cases, the route is not usable so communications should not continue.