Communication Basics #32 Worked Example: STM-1 + ATM Cells — Frame Efficiency, 83.17% Real Data
Transmission Lines and Network ConnectionsInstructor: Dr. Süleyman Burak ÇELİK
We open up an SDH frame and ask a very practical question — how much of it actually carries user data, and how much is wasted on framing overhead. The answer for STM-1 filled with ATM cells is 83.17%, and the video walks every step of how that number falls out. The STM-1 frame is a 9 × 270 byte grid — 2,430 bytes, 19,440 bits per frame. Of those, the first 9 columns are section overhead (RSOH + AU pointer + MSOH = 81 bytes), 3 columns are path overhead (POH = 27 bytes), and 258 columns × 9 rows = 2,322 bytes form the payload area we get to fill. STM-1 is transmitted at 8,000 frames per second, so each frame lasts exactly 125 microseconds — the heartbeat of SDH. Then the cell. Each ATM cell is 53 bytes: 5 bytes of header (routing, traffic class, error check), 1 byte of AAL adaptation overhead inside the 48-byte payload, leaving 47 bytes of actual user data. Cell-level efficiency: 47/53 ≈ 88.7% — already 11% overhead just at the cell level. Now the math. Cells are atomic, but SDH allows them to span two consecutive payload rows, so we round 2,322 / 53 = 43.81 down to **43 cells per frame**. Each cell carries 47 × 8 = 376 user bits, so 43 × 376 = **16,168 user bits per frame**. Efficiency: 16,168 / 19,440 ≈ **83.17%**. Data rate: 16,168 bits / 125 µs = **129.344 Mbps** of real user data, against the STM-1 line rate of 155.52 Mbps. Roughly 26 Mbps of every link goes to overhead — framing structure, cell headers, and the unused tail of the payload. The closing third explains why real data never equals line rate. Three takeaways: (1) framing overhead is a tax — every protocol layer adds bytes the user does not see; (2) small fixed cells like ATM trade efficiency for predictability — voice cannot be held up by a giant data packet, but the small cell size means a relatively large header-to-payload ratio; (3) the same STM-1 frame with different fillings gives dramatically different efficiency — next videos run the same calculation with 576-byte IP packets and a mixed payload. Topics covered: - STM-1 frame anatomy (9 × 270 bytes, RSOH/AU/MSOH/POH overhead, payload area) - 8,000 frames/s = 125 µs per frame (the SDH heartbeat) - ATM cell anatomy: 5-byte header + 1-byte AAL + 47-byte user data = 53 bytes - atomic cell packing across two payload rows: floor(2,322/53) = 43 cells - user data per frame: 47 × 8 × 43 = 16,168 bits - frame efficiency 83.17% and where the missing 17% goes - user data rate 129.344 Mbps vs line rate 155.52 Mbps (≈ 26 Mbps overhead) - why ATM chose small fixed cells (predictable voice latency) Sources: ECE 271 ExQ31 == HW Q44. Part of the AcEdumy Communication Basics series.