Complementary encoder of stealth signal is proposed and demonstrated for coding, modulating and enhancing the privacy of optical stealth transmission. With complementary encoding, the stealth signal carried by amplified spontaneous emission (ASE) light keeps the same characteristic to ASE noise and can be concealed well under public channel. The experiment results demonstrate the feasibility of the scheme and show the stealth signal has the same impact on public channel in transmission performance, compared to the ASE noise.
© 2014 Optical Society of America
With the development of optical networks, it is more important to promise the optical communication networks are secured. Variously approaches to enhance the security in optical networks had been researched [1–4]. Dissimilar to quantum private communication and chaos encryption, optical steganography build a private channel over the existed public channel, making it undetectable [5, 6]. In previous optical steganography methods, to get a parallel characteristic with the noise in optical networks, most attention are paid on stretching the optical spectrum and decreasing the peak power of optical pulse [5–8]. Recently, an optical steganography scheme based on Mach-Zehnder interferometer (MZI) has been proposed and demonstrated which use amplified spontaneous emission (ASE) noise as a carrier to transmit data signal [9, 10].
In this paper, we propose a complementary encoder based on wavelength selection switch (WSS). With the complementary encoder, we demonstrate the transmission of optical stealth signal hidden under a public wavelength-diversion multiplexing (WDM) network. In this optical stealth transmission system, stealth signal is generated by an ASE light source which have the same characteristic with the ASE noise in the optical networks. Being complementary encoded, the ASE signal keep the characteristic undifferentiated and is covered by public signal effectively both in temporal and spectral domain after combing with public channel. By correctly decoding, error free detection of stealth channel can be achieved over 40 km single model fiber (SMF).
2. Complementary coding optical steganopgraphy
In the optical stealth communication system, to improve the private in temporal domain for stealth channel we should try to stretch the optical pulse through optical code division multiple access (OCDMA) encoder or chromatic dispersion device. However, the bandwidth for optical signal is difficult to change unless through nonlinear processing. Thus, it is a good approach to choose an optical source with widely bandwidth for enhancing security in spectral domain of stealth channel. ASE light source based optical stealth scheme can feet thus demand both in temporal and spectral domain [9, 10]. In our optical stealth transmission system, we employ complementary coding and modulation scheme without changing the characteristic of the ASE light.
We experimentally demonstrate the complementary coding and modulation scheme. Figure 1 shows the setup of complementary encoder in the optical stealth transmission system. An ASE light is injected to a WSS which is on 50 GHz ITU-T grid and has 96 channels in total. The WSS is a maturely commercial device, and can easily reassign the channel in a short time less than one second. The first port of WSS is encoded with one code which we define it as Code and the second port is encoded with the complementary code which we define it as Code*. The encoded signal by Code is modulated by intensity modulator using 27-1 pseudo-random binary sequence (PRBS) at 2.5 Gb/s while the complementary encoded signal is modulated by another intensity modulator using the opposite 2.5 Gb/s PRBS. An optical tunable delay line (OTDL) is selected to adjust the delay for signal of the two ports. In addition, two variable optical attentions (VOA) are used to equalize the optical average power of the two signals. Two path signals are coupled by a 50:50 optical coupler.
The measured results of the ASE signal both in temporal and spectral domain are shown in Fig. 2. As shown in Fig. 2(a), two opposite codes are used in the experiment and the combined signal has the similar spectrum with ASE signal. Figures 2(b) and 2(c) show the waveform of encoded signal and complementary encoded signal, while Figs. 2(d) and 2(e) are the waveform of original and combined signal. Thus, the complementary coding and modulation do not transform the characteristic ASE signal in temporal and spectral domain. After 40 km transmission in single model fiber (SMF), Fig. 2(f) shows the waveform of decoded signal by a WSS with the same code to Code. Influenced by the inexactitude of dispersion compensating, Figs. 2(b) and 2(f) have less difference.
Thus, the complementary encoder completes both the coding and modulation for ASE signal without changing its characteristic. We should note that the WSS is programmable. The code corresponds to the wavelength, “1” represents the corresponding wavelength is turned on while “0” is turned off.
3. Experimental setup
Based on complementary coding, the schematic diagram of the optical stealth transmission system is shown in Fig. 3. The public WDM signal is generated by a directly modulated DFB laser module at 1551.72 nm and the modulating data is 215-1 PRBS at 2.5 Gb/s. For the stealth channel, the optical signal is created by ASE light source with a center wavelength of 1546.4 nm and a 20 dB bandwidth of 45.2 nm. The stealth signal is coded and modulated by the complementary encoder. An additional ASE light source is added to simulate the ASE noise in the optical networks. The stealth signal is injected into the public channel through a 50:50 optical coupler. The combined signal of public and stealth channel sent over a 40 km SMF followed by a + 676 ps/nm dispersion compensation fiber (DCF). At the receiver side, a 50:50 optical coupler splits the combined signal to two ports and the public channel receives the transmitted signal output from the first port. The stealth channel receives the signal output from WSS which holds the same code with the complementary encoder and turns off the wavelength channel corresponding to public signal.
In the experiment, the public and stealth receivers are made up of photo-detector module and clock data recovery module. For stealth channel, we use a band pass filter from 1559.05 to 1579.05 nm replaced the WSS in the receiver and it stands for one code which has a piece of all “1” during that wavelength spacing.
4. Results and discussion
4.1 Temporal and spectral domain measurement
The measured eye diagrams and spectrum diagrams show the stealth channel has hidden effectively both in temporal and spectral domain. The eye diagrams of public channel without and with stealth channel are shown in Figs. 4(a) and 4(b), while the optical power of public signal is 11.2 dB more than stealth. The comparability of the two eye diagrams explain the stealth channel has covered by the public channel in temporal domain. Figure 4(c) shows the spectrum of public channel with stealth channel and with ASE noise. The resemblance of the two spectrums indicates that the stealth channel has been hidden under public channel in spectral domain. The peak power of the public channel is 34.2 dB higher than the stealth channel.
If the public channel is under attacking, the eavesdropper can demodulate the public signal easily. With the analysis of the public signal in temporal and spectral domain, the eavesdropper cannot detect the existence of stealth channel.
4.2 Power penalty of the system
Figure 5 shows the measured results of bit error rate (BER) tests. The measured results show that combing the public channel results in 0.68 dB power penalty to the stealth channel. In Fig. 5(a), error-free transmission for stealth channel has been achieved with 40 km SMF. While the power of stealth signal is 11.2 dB less than the public, the combing of public signal will cause 0.68 dB power penalty at 1 × 10−9 BER. This is because the amplifying of public signal results in ASE noise. With the decreasing of the optical power in stealth channel, the impact of the ASE noise is added and the BER adds to more than 4 × 10−9. A better low noise amplifier or accurate dispersion compensator can improve the performance of stealth channel.
For public channel, the stealth signal causes the same power penalty as the additional ASE noise. In Fig. 5(b), the BER curves of the public channel with noise and stealth signal are indistinguishable. It is owing to that the two signals have the same characteristic in temporal and spectral domain. The combing of stealth signal only causes 0.1 dB power penalty for public channel at 1 × 10−9 BER.
4.3 Code space and security analysis
The measured results show that the stealth channel is safety to avoid the eavesdropper’s discovering the presence of an additional channel except the public channel. However, if the eavesdropper knows this optical stealth scheme or suppose there is a covert channel. Whether the stealth can be detect should be discussed when she or he is trying to recover the covert signal by the similar device to the system. In the following, we will experimentally analyze the security for the stealth channel against an eavesdropper who knows the principle of coding, but is unknown with the correct code.
In this optical stealth transmission system, the code length of the complementary encoder is even and the code weight is the half of code length. The WSS has 96 channels and it is covered by the wavelength spacing of ASE signal. Therefore, the maximum code length is 96 and the code space is 6.4 × 1027. However, the code space is not equal to the security of the stealth channel. If the dispersion has been accurate compensated, the eavesdropper who holds the same WSS and an EDFA with low noise can brute force cracking the codes. With a proper proportion of right chips, the eavesdropper can recover the waveform of stealth signal or complementary stealth signal. Thus, the eavesdropper can confirm the presence of a covert channel and detect the stealth signal.
We now consider the attacking by the eavesdropper holds a matching dispersion compensator, a WSS same as the stealth channel and an EDFA. The public channel corresponds to only one channel in the WSS, so we ignore the affection by it in the following discussion. To find the code of the stealth channel, the eavesdropper turns on 48 channels in the WSS to receive all the chips of Code and eliminate the chips of Code*. The received signal is consisted by m chips of Code and 48-m chips of Code*. The corresponding distributed probability is:Fig. 6. The sum of P(m) is 0.976 while R1 = m/48 ranges from 0.4 to 0.6.
We do an experiment to attack the stealth channel in the proposed system. The eavesdropper holds a WSS, an EDFA and a matching dispersion compensator. While the eavesdropper attack a single code chip which equal to a 0.4 nm width wavelength in this system, the eavesdropper cannot detect the stealth signal. It is because that the power is too small to be detected even though using an EDFA with low noise. When the WSS turns on 20 channels for searching the chips of Code, we define the proportion of the chips of Code during the 20 channels is R’1. Figures 7(a)-7(c) show the waveform of received signal with different R’1. When R’1 is 1.0, the chips received by the eavesdropper are all belong to Code. With the decreasing of R’1, the detection of the stealth signal goes difficult. When R’1 is more than 0.6, the stealth signal can be discovered by analyzing the waveform. Figure 7(d) shows the BER curve with different R’1. An error free detection can be achieved when R’1 is larger than 0.9. While R’1 is less than 0.7, the BER tester loss the synchronization and cannot detect the stealth signal. We should note that when R’1 is less than 0.4, the received chips are mainly belong to Code*, and the eavesdropper can discover the waveform of the complementary stealth signal. Thus, the stealth channel cannot be detected with a confidence level of 97.6%.
To improve the security of the stealth channel, we can add the power of additional ASE noise. High-power ASE noise will degrade the probability for the stealth channel of being intercepted. However, it also degrades the transmission performance of both the public channel and stealth channel. As another approach, we can insert multi-users in the stealth channel. Multi-users in the stealth channel can increase the difficulty of interception .
4.4 Further discussion
One limitation to the propagation length of the system is the dispersion. The 20 dB bandwidth of the stealth signal is 45.2 nm which is much larger than the public signal. So it is a key issue to compensate the dispersion accurately for the stealth channel. On the other hand, it is the gain competition effects in EDFA who limits the stealth signal to transmit over long distances. When an EDFA is used to amplify the signal in the long-distance transmission system, the public channel can be effectively amplified. However, the stealth channel and the ASE noise cannot be amplified effectively owing to the gain competition effects. The stealth signal cannot transmit much longer after the EDFA.
If the stealth signal is transmitted over a WDM system with multiple channels at different wavelengths, the proposed approach can also work. The 20 dB bandwidth of public channel shown in Fig. 4(c) is 0.3 nm which is equal to 37.5 GHz and can be contained by one channel of the WSS. For a WDM system with multiple channels on the 100 GHz ITU-T grid, the public signal can be eliminated by turning off the matching channels in the WSS. The residual channels of the WSS can be used for recovering the stealth signal.
A complementary coding and modulation scheme has been proposed and experimentally demonstrated using WSS and ASE light source. We experimentally demonstrated the optical steganography based on this scheme in the following and it takes the advantage of ASE signal’s good stealth security in optical networks. The combing of stealth signal results in only 0.1 dB power penalty to public channel, while the coupling with public signal causes 0.68 dB power penalty to stealth channel. In the experiment, the stealth signal shows the same characteristic with ASE noise in temporal and spectral domain and same impact to public channel.
The work described in this paper was supported in part by the National Natural Science Foundation of China under Grants with No. 61475193, No.61177065, No.61174199, and the Jiangsu Province Science Foundation Council with No. BK2012058. We are also obliged to the reviewers for the careful review and helpful comments.
References and links
1. A. Argyris, D. Syvridis, L. Larger, V. Annovazzi-Lodi, P. Colet, I. Fischer, J. García-Ojalvo, C. R. Mirasso, L. Pesquera, and K. A. Shore, “Chaos-based communications at high bit rates using commercial fibre-optic links,” Nature 438(7066), 343–346 (2005). [CrossRef] [PubMed]
2. N. Q. Li, W. Pan, L. S. Yan, L. Bin, M. F. Xu, Y. L. Tang, N. Jiang, S. Y. Xiang, and Q. Zhang, “Chaotic optical cryptographic communication using a three-semiconductor-laser scheme,” J. Opt. Soc. Am. B 29(1), 101–108 (2012). [CrossRef]
4. L. L. Yi, T. Zhang, Z. X. Li, Y. Zhang, Y. Dong, and W. S. Hu, “Power-penalty-free all-optical decryption using stimulated Brillouin scattering in optical fiber,” Laser Phys. Lett. 10(4), 045102 (2013). [CrossRef]
5. K. Kravtsov, B. Wu, I. Glesk, P. R. Prucnal, and E. Narimanov, “Stealth transmission over a WDM network with detection based on an all-optical thresholder,” IEEE/LEOS Annual Meet. 480–481 (2007). [CrossRef]
6. M. P. Fok and P. R. Prucnal, “Compact and low-latency scheme for optical steganography using chirped fiber Bragg gratings,” Electron. Lett. 45(3), 179–180 (2009). [CrossRef]
7. Z. Wang and P. R. Prucnal, “Optical steganography over a public DPSK channel with asynchronous detection,” IEEE Photon. Technol. Lett. 23(1), 48–50 (2011). [CrossRef]
8. Z. Gao, X. Wang, N. Kataoka, and N. Wada, “Stealth transmission of time domain spectral phase encoded OCDMA signal over public WDM network,” IEEE Photon. Technol. Lett. 22(13), 993–995 (2010). [CrossRef]
9. B. Wu, Z. X. Wang, Y. Tian, M. P. Fok, B. J. Shastri, D. R. Kanoff, and P. R. Prucnal, “Optical steganography based on amplified spontaneous emission noise,” Opt. Express 21(2), 2065–2071 (2013). [CrossRef] [PubMed]
10. B. Wu, Z. Wang, B. J. Shastri, M. P. Chang, N. A. Frost, and P. R. Prucnal, “Temporal phase mask encrypted optical steganography carried by amplified spontaneous emission noise,” Opt. Express 22(1), 954–961 (2014). [CrossRef] [PubMed]
11. T. H. Shake, “Security performance of optical CDMA against eavesdropping,” J. Lightwave Technol. 23(2), 655–670 (2005). [CrossRef]