Distinguishing services at physical layer level in autonomic optical Internet

This paper proposes a mathematical model for distinguishing services at bit-level. By using correlation analysis, the features of bit flows are extracted. The feature analyses of time division multiplex (TDM) service and Gigabit Ethernet (GE) service are described in detail, and then the performance of the proposed method is evaluated. Both of the undetected probability and the misjudgment probability for GE service are as low as 2% at the average Ethernet frame length of 300 B.