Error Control and flow control in TCP/IP?

[Diellza]

Hello friends , I too have a query regarding the the networking, I am a networking student , I have exams in the upcoming week , my Networks subject is very weak and I have already failed in one of the unit but now I wish to clear it, so I would appreciate if you could reply to my query, my question is that how is error and flow control done in TCP/IP protocol. I am expecting your replies as early as possible as I ave hot exams on my head.

[Amie]

TCP is a stateful protocol that provides the a lot of features such as reliability , error control and flow control of the the packet, the error control implies that it checks that the packet is not corrupted or is not lost in the transaction. The error is mostly checked after calculating value of the checksum , sum of the checksum being derived and the packets' value must be equal to zero, if it is not then the packet is discarded.

[BiGMaCHiNe]

Flow control is a method whose main reason is to correctly match the transmission rate of dispatcher to that of the recipient and the network. It is significant for the transmission to be at a high enough rate to guarantee superior performance, but also to guard against irresistible networks Just note flow control is not as same as congestion control. Congestion control is mainly concerned with a constant excess of network middle devices such as IP routers.

[HiSpeed]

In case of the error and flow control , if the sender sends any packet is lost and then the recipient does not receive any packet, so it will be quite obvious that it will not send any acknowledgment so in order to do that sender will need to set a timer after particular time out it will consider the packet as lost and then will retransmit the packet, for error control I think Sliding windows concept was introduced.

[Lemog]

Modern data networks are intended to support a different range of hosts and communication channels. Take an example of a 200MHz Pentium-based host sending packets to a 25MHz 80386/SX. Clearly, the Pentium will be capable to sink the slower processor with packet. Similarly, think about two hosts, both utilizing an Ethernet LAN, but with the two Ethernets linked by a 28.10 Kbps modem connection. If one machine starts sending to the other at Ethernet velocity , the modem connection will rapidly become inundated. In both scenarios , flow control is required to speed up the data transfer at an satisfactory speed.

[Carley]

Request/reply flow control need each data packet to be recognized by the distant host before the subsequent packet is transferred. Sliding window concept which is being utilized by TCP, allow several data packets to be in immediate transfer, making more proficient utilization of network bandwidth. Lastly, Internet's Unreliable Delivery Model permits packets to be rejected if network resources are not accessible, and asks the protocols make arrangements to retransmit the packets again.