Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

I have a question about the limitations on the segment size in the transport layer.

Since the data has to get to the lower layers why will it be fragmented? Finally the data is sent in bits at the physical layer.

Why do we bother about segment length in transport layer when we have another layer to take care of it?

share|improve this question
Probably because too large segments would cause fragmentation at the IP layer, thus decreasing performance. en.wikipedia.org/wiki/IP_fragmentation –  slhck Jan 6 '13 at 15:00
okay is the size of the segment is large then it is time consuming for the ip layer to fragment it into packets ... ?? –  CollegeStudent Jan 6 '13 at 15:07
It's not time consuming per se, but error-prone. IP packets may be lost or arrive late. The more IP packets you need to transmit a TCP segment that in case of error would have to be retransmitted, the worse the performance. –  slhck Jan 6 '13 at 15:09
okay ... i get it ..one more thing we segment the data in the transport layer and those segemnets are sent to the n/w layer and again it is fragmented into packets ... and sent to the datalink layer whr it gets fragmented and sent to the destination ... but what is meant by lossing a frame ... is it loosing streams of bits associated with that frame ?? and could be found out only when it gets defragmented ? –  CollegeStudent Jan 6 '13 at 15:17
A TCP packet contains sequence numbers and a 16-bit checksum field. This allows you to check for header or payload errors. Although a rather weak error detection mechanism, it works quite well. I'd wait for someone more knowledgeable than me in networking to post a real answer though. –  slhck Jan 6 '13 at 15:20

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.