Protocol Functions
The IP provides a number of core functions and associated procedures to carry out the various harmonizing functions that are necessary when internetworking across dissimilar networks. These include the following:
(1) Fragmentation and reassembly:    This concerns the transfer of user messages across networks/subnets which support smaller packet sizes than the user data.
(2) Routing: To perform the routing function, the IP in each source host must know the location of the internet gateway or local router that is attached to the same network or subnet. Also, the IP in each gateway must know the route to be followed to reach other networks or subnets.
(3) Error reporting: When routing or reassembling datagrams within a host or gateway, the IP may discard some datagrams. This function is concerned with reporting such occurrences back to the IP in the source host and with a number of other reporting functions.
  Internet Control Message Protocol
Internet Control Message Protocol (ICMP) forms an integral part of all IP implementations. It is used by both hosts and gateways for a variety of functions; especially by network management. The main functions associated with the ICMP are as follows:
Error reporting Reachability testing Congestion control
Route change notification Performance measuring         Subnet addressing.
The message types associated with each of these functions are shown in Table 10.2. Since the IP is a best-try (unacknowledged) protocol, datagram may be discarded while they are in transit across the internet. Of course, transmission errors are one cause, but datagram can be discarded by a host or gateway for a variety of reasons. In the absence of any error reporting functions, a host does not know whether the repeated failure to send a datagram to a given destination is the result of a poor transmission line (or other fault within a network) or simply the destination host being switched off. The various messages associated with the error reporting function are used for this purpose.
If the datagram is corrupted by transmission errors, it is simply discarded. If a datagram is discarded for any other reason, the ICMP in the host or gateway that

discards the datagram generates a destination unreachable error report message and returns it to the ICNIP in the source host with a reason code. Reasons include the following:
Destination network unreachable. s  Destination host unreachable. Specified protocol not present at destination.
Fragmentation needed but don't fragment (DF) flag set in datagram header. Communication with the destination network not allowed for administrative reasons.
Communication with the destination host not allowed for administrative reasons.
Other error report messages include time exceeded, which indicates that the time to live parameter in a discarded datagram has expired and parameter error, which indicates that a parameter in the header of the discarded datagram was recognized.
If the network manager receives reports from a user that a specified destination is not responding, the reason must be determined using the reachability testing function. Typically on receipt of such a report, the network manager initiates the sending of an echo request message to the support host to determine whether it is switched on and responding to requests. On receipt of an echo request message, the ICMP in the destination simply changes this to an echo reply message, which it returns to the source. A similar test can be performed on selected gateways if necessary.
If the datagram is discarded because no free memory buffers are available as a result of a temporary overload condition, a source quench message is returned to the ICMP in the source host. Such messages can be generated either by host or by a gateway. They request the source host to reduce the rate at which it sends datagrams. When a host receives such a message, it reduces the sending rate by an agreed amount. A new source quench message is generated each time a datagram is discarded so that the source host incrementally reduces the sending rate. Such messages help alleviate congestion control within the Internet.
When the network has multiple gateways attached to it, a gateway may receive datagrams from a host even through determines from its routing table that they would be better sent via a different gateway attached to the same network. To inform the source host of this, the ICMP in the gateway returns a redirect message to the ICMP in the source indicating which is the better gateway to the specified destination. The ICMP in the source then makes an entry in its routing table for this destination.
An important operational parameter for an internet is the mean transit delay of datagrams. This is a measure of the time a datagram takes to traverse the Internet from a specified source to a specified destination. To ascertain this time, a host or a network manager can send a timestamp request message to a specific destination. Each message contains the following three time-related parameters (which constitute the timestamp).
(1) The time the datagram was sent by the source.
(2) The time the datagram was received by the destination. (3) The time the datagram was returned by the destination.
On receipt of a timestamp request message, the ICMP in the destination simply fills in the appropriate timestamp fields and returns the datagram to the source. On receipt of the reply, the source can quantify the current round trip delay to that destination and from this determine the datagram transit. delay.
Finally, when subnet addressing is being used, the address mask request and corresponding reply messages are used by a host to ascertain the address mask associated with a local subnet. This is needed by a host to determine, for example, whether a specified destination is attached to the same subnet. The address mask is held by the local router associated with the subnet. The ICMP in a host can obtain the address mask sending a request message and reading the mask from the reply.

  SHORTCOMINGS OF IPv4

The following are some major drawbacks of IPv4.
(1) With the 32-bit addresses, IPv4 currently allows for approximately 4 billion IP addresses. Although during the time of its creation IPv4 seemed to have enough addresses, the exponential growth of Internet has pressed for more addresses.(2) The Internet has grown with the addition of more and more networks and sub-networks. Routers, which route data through these networks, must store up-to-date information of all these sub-networks. The IP addresses, which are stored in routing tables had to be divided into various classes to accommodate every sub-network. Hence the size of the routing tables has grown along with the Internet and now this poses a problem of decreasing the performance of the Internet.
(3) Security is a very important factor in the globally connected Internet. IPv4 does not provide any security mechanism such as the authentication of the packet sender or the encryption of the packets before delivery.
(4) IPv4 does not have any feature to know the kind of data transmitted. This knowledge can be crucial especially when we are dealing with real-time or multimedia data that should be treated as high priority data.
(5) Another problem with an IPv4 network is of "lost packets" within a network. The TTL field in the IP header sets the lifespan of a datagram. If the datagram is unable to reach the destination within this time, it will expire. The computer on the receiving side simply requests that the datagram be resent. This process can become time-consuming and cause problems especially if we are dealing with real time data.
(6) During the process of broadcasting, IPv4 sends a packet to every address on the network. Some uses, routers and other network resources tend to become overloaded because of this. This insecure process also can cause network congestion and excessive collision.
  IP NEXT GENERATION
The rapid expansion in the number of interconnected networks making up the Internet means that 32-bit addresses currently used with IPv4 will need to be extended in the not too distant future if the current rate of growth continues. In anticipation of this, the Internet Engineering Task Force (IETF) has embarked upon the specification of a successor to the current IPv-4 protocol. It is known as IP next generation (IPng) or more correctly, IP version 6 (IPv6).

BACK