sequence number in TCP

Question

Why do we need the sequence number and the next sequence number field in the TCP header?

Below is a TCP header from a packet captured using wireshark.

Answer 1

First, fields in Wireshark enclosed by [brackets] are computed fields - they're not in the packet. That next sequence number field shown by Wireshark is one such field. Wireshark is computing that by taking the 'sequence number' field and adding it to the payload size of your packet. It's no surprise then that the difference between these two numbers is 1430 - a common TCP payload size.

Sequence numbers in TCP are in units of bytes - they basically say, what byte location in the TCP stream this packet's payload is inserted at.

The 'acknowledged' sequence number shows how many bytes I'm acknowledging as having received.

Since TCP is bidirectional, each end has to declare

1. Where the bytes its transmitting should go in the stream and
2. What bytes that you've sent me that I've received.

As such, each TCP packet has two fields that refer to sequence numbers - the 'sequence number' field, and the 'acknowledgment number' field.

Without the 'sequence number' field, the receiving end wouldn't be able to tell if packets were received out of order. Without the 'acknowledgment number' field, the transmitting end wouldn't know if some of his packets had been dropped and the receiver never received them.