Implement TCP Data Transfer with Sequence Numbers and ACKs

Topic Overview

Sequence Numbers and ACKs

Once TCP is established, data flows as a reliable byte stream. Each byte gets a sequence number. For example, sending 100 bytes starting at SEQ 1000 means the receiver ACKs with 1100 (the next expected byte).

Data Transfer Example

• Segment 1: SEQ=1000, 100 bytes of payload
• Receiver ACKs: ACK=1100
• Segment 2: SEQ=1100, next 100 bytes
• Receiver ACKs: ACK=1200

If segments arrive out of order (SEQ 1200 before 1100), the receiver buffers the gap and waits. Duplicate ACKs for the last in-order byte (three in a row) trigger fast retransmit.

Segment Limits

A TCP segment is header (20+ bytes) plus payload up to MSS. On Ethernet, MSS is typically 1460 bytes (1500 MTU minus 20-byte IP header minus 20-byte TCP header). The sender cannot exceed the receiver's advertised window.

Why This Exercise

This task asks you to simulate data transfer with printed SEQ/ACK pairs. The Linux kernel's tcp_sock struct maintains this same bookkeeping per connection, and Wireshark's "Follow TCP Stream" reconstructs it from captured segments. Production stalls from Nagle's algorithm interacting with delayed ACKs trace directly back to misunderstanding this sequence-number accounting on port 80 and other TCP services.

Given a starting sequence number, a total byte count, and an MSS, emit the segment layout of the transfer and the receiver's final cumulative ACK. Sequence numbers advance by payload bytes only (RFC 9293): a segment carrying N bytes consumes N sequence numbers, so the next segment starts at seq + N, and the final ACK equals initial_seq + total_bytes.

Input (stdin): one line: initial_seq total_bytes mss (all decimal, 64-bit-safe).

Output (stdout):

• one line per segment: SEG seq=<s> len=<l>, where len is min(mss, bytes remaining) and each next seq is the previous seq + len
• final_ack=<initial_seq + total_bytes> - the next byte the receiver expects (the cumulative ACK number)
• a zero-byte transfer emits no SEG lines and final_ack equals the initial sequence number