256b/257b Transcoding for 100 Gb/s Backplane and Copper Cable

20

IBM

Benefits of 256b/257b TC vs. 512b/514b TC

The granularity of 256b/257b TC simplifies processing of alignment markers because the total number

of 20 alignment markers is divisible by 4 (# of control/data blocks in 256b-based TC) but not by 8 (# of

control/data blocks in 512b-based TC): gustlin_01_0312.pdf

RS(528,514), t=7, m=10 code provides the same error-rate performance and the same coding gain of

4.87 dB in case of 256b/257b or 512b/514b TC scheme

5% to 7.5% reduction in overall latency associated with transcoding and FEC

15% reduction in total gate complexity

15% reduction in total power dissipation

FEC Code

RS(n, k, t, m)

Transcoding

(TC)

Effective Gain

Overall

Latency *

Total Area

(40nm gates) *

Total

Power *

RS(528, 514, 7, 10)

256b/257b

4.87 dB

94.3 ns

244k

90 mW

RS(528, 514, 7, 10)

512b/514b

4.87 dB

99.4 ns

285k

105 mW

gustlin_01_0112.pdf and gustlin_01_0312.pdf

* Zhongfeng Wang computed overall latency, total gate complexity and power dissipation for a nominal design