Communication Basics #33 Worked Example: STM-1 + 576-byte IP Packets — 90.87% Real Data

Instructor: Dr. Süleyman Burak ÇELİK

Same STM-1 frame, different filling. In the previous video we filled the payload with 53-byte ATM cells and got 83.17% efficiency. Today we replace those tiny cells with 576-byte IP packets — and watch efficiency climb to 90.87%, with user data rate jumping from 129.344 Mbps to 141.312 Mbps. Twelve more megabits per second of useful data through the same 155.52 Mbps pipe, just by changing what we put inside the frame. The IP packet is 576 bytes total — historically the default safe size from RFC 791. Of those, 24 bytes are header overhead (IP header plus protocol stack allowance) and 552 bytes carry actual user data. Packet-level efficiency: 552/576 ≈ 95.83%, vastly better than ATM's 47/53 ≈ 88.7%. Why? The header is fixed regardless of payload size. The bigger the packet, the smaller the header looks proportionally. Same idea every time: bigger packets amortize the header. Now the packing math. The STM-1 payload area is still 2 322 bytes — that number does not change. 2 322 / 576 = 4.031, round down to 4 complete IP packets per frame. Used: 4 × 576 = 2 304 bytes. Tail: only 18 bytes unused, versus 43 wasted with ATM. Bigger pieces mean fewer pieces mean less tail waste. Each packet carries 552 × 8 = 4 416 user bits, so 4 × 4 416 = 17 664 user bits per frame. Efficiency: 17 664 / 19 440 ≈ 90.87% — a 7.7-point improvement over ATM. Data rate: 17 664 / 125 µs = 141.312 Mbps. The closing third is about the tradeoff. Bigger packets win efficiency but lose latency predictability. A 1 500-byte IP packet hogging the line means a voice call has to wait its turn. ATM chose tiny fixed cells specifically to keep voice latency low; IP chose larger variable packets specifically because data flows benefit from amortizing the header. Three takeaways: (1) bigger packets amortize the header — same 24 bytes spread over more user data; (2) the unused tail measures the size match — 18 bytes with 576B IP versus 43 with ATM; (3) efficiency comes at the cost of latency predictability — the small-cell choice is a service-class decision, not just an engineering one. Topics covered: - STM-1 frame recap (9 × 270 = 2 430 bytes, 19 440 bits, payload area 2 322 bytes) - IP packet anatomy: 24-byte header + 552-byte user data = 576 bytes - packet-level efficiency 95.83% (vs ATM cell 88.7%) - packet packing: floor(2 322 / 576) = 4 packets, 18-byte tail - user data per frame: 552 × 8 × 4 = 17 664 bits - frame efficiency 90.87% — a 7.7-point improvement over ATM - user data rate 141.312 Mbps vs ATM's 129.344 Mbps (≈ 12 Mbps more useful data) - the bigger-packets-vs-latency tradeoff (why ATM still chose tiny cells) Sources: ECE 271 ExQ38 == HW Q45. Part of the AcEdumy Communication Basics series.