CN107817967B

CN107817967B - SFP (Small form-factor pluggable) based integrated quantum random number generator

Info

Publication number: CN107817967B
Application number: CN201711063647.9A
Authority: CN
Inventors: 富尧; 朱伟
Original assignee: Zhejiang Shenzhou Liangzi Network Science & Technology Co ltd
Current assignee: Zhejiang Shenzhou Liangzi Network Science & Technology Co ltd
Priority date: 2017-11-02
Filing date: 2017-11-02
Publication date: 2024-04-12
Anticipated expiration: 2037-11-02
Also published as: CN107817967A

Abstract

The invention discloses a quantum random number generator based on SFP receiving and transmitting integration, which comprises: the SFP is provided with an emergent end for transmitting optical signals, an incident end for receiving the optical signals and an output end for outputting digital signals; the interference device is used for receiving the optical signals at the SFP emergent end, transmitting the optical signals to the SFP incident end after interference, and converting the randomness of the laser phase into the randomness of the intensity; and the FPGA is used for carrying out digital acquisition and post-processing on the digital signal at the SFP output end and finally generating the quantum random number. In the invention, a light source part and a detection part after interference in the prior art are integrated, an SFP with an integrated structure is adopted, a DFB laser light source arranged in the SFP is utilized to emit light signals, the light signals are received by an incidence end of the SFP after interference, corresponding photoelectric conversion and analog-to-digital conversion processing are carried out in the SFP, and then digital signals are sent to an FPGA through an output end; the product structure is simplified on the whole, and the integration level is improved and the volume is reduced.

Description

SFP (Small form-factor pluggable) based integrated quantum random number generator

Technical Field

The invention relates to the technical field of quantum communication, in particular to a quantum random number generator based on SFP (small form-factor pluggable) transceiver.

Background

Random numbers have important applications in many areas of modern society, and in a series of important areas such as authentication, scientific computing, and cryptographic systems, high-quality random numbers are indispensable. Existing mainstream random number generation schemes are based on computer software that generates a new series of random numbers by a mathematical function on the basis of a seed file (a set of random numbers). Such random numbers are generally referred to as pseudo-random numbers because they are not of any randomness in nature. The wide range of applications of pseudo-random numbers benefit from their almost zero cost and the associated procedures can be easily applied to a variety of practical application scenarios. But the application of pseudo-random numbers in any security-related field will face a great risk, especially in the age of today's computer computing capabilities fast-moving, due to its inherent non-randomness. Therefore, the application of true random numbers in the security field greatly improves the security of the system. Traditionally, removing the above-mentioned pseudo random numbers can also be achieved by measuring classical physical noise sources to obtain random numbers, such as circuit thermal noise. However, the classical physical theory does not have any randomness, and the randomness of the classical physical noise source is due to the complexity of the system and not the intrinsic randomness, so the random number based on this cannot be called true random number. Moreover, the generation rate of the physical source random number is slow, and the statistical property is poor, so that the physical source random number is never used on a large scale.

In physics, quantum mechanics theory is intrinsic and contains randomness, so that true random numbers can be obtained by measuring quantum physics phenomena, which is the only way to obtain the true random numbers at present. Random number generators based on quantum physics are generally referred to as quantum random number generators, and since they have been proposed, they have been developed continuously. Quantum random number generators based on laser phase random fluctuations are considered to be the closest to the marketed quantum random number scheme. The scheme is simple, the adopted devices are all universal optical devices, so that the cost is economical, the random number generation rate can be up to tens of Gbps, and the system is stable and reliable.

Based on security considerations, it would be a necessary trend for pseudo-random numbers to be replaced by quantum random numbers. Referring to fig. 1, the working process of the quantum random number generator based on laser phase random fluctuation in the prior art is as follows:

1. an external DFB laser emits continuous laser light.

2. The laser enters the interference device to interfere.

3. And measuring the interfered laser by utilizing a photoelectric detection module.

And 4, the data acquisition module of the FPGA carries out digital sampling on the measured electric signals to obtain original random numbers.

5. The original random number is processed by a post-processing module to obtain the quantum random number.

6. The generated quantum random number is output to the corresponding random number using device through the port.

The quantum random number generator is an important basic device in the field of quantum information, and since the technology is an emerging field, the main research and development and attention directions are the stability of the operation of the quantum random number generator and the generation rate of the quantum random number, but related researches are lacking in the aspects of structure simplification, cost reduction and miniaturization, which prevent the further popularization and application of the quantum random number generator.

Disclosure of Invention

The invention provides a quantum random number generator with a simplified structure and a smaller volume.

A quantum random number generator based on SFP transceiving integration, comprising:

the SFP is provided with an emergent end for transmitting optical signals, an incident end for receiving the optical signals and an output end for outputting digital signals;

the interference device is used for receiving the optical signals at the SFP emergent end, transmitting the optical signals to the SFP incident end after interference, and converting the randomness of the laser phase into the randomness of the intensity;

and the FPGA is used for carrying out digital acquisition and post-processing on the digital signal at the SFP output end and finally generating the quantum random number.

In the invention, a light source part and a detection part after interference in the prior art are integrated, an SFP with an integrated structure is adopted, a DFB laser light source arranged in the SFP is utilized to emit light signals, the light signals are received by an incidence end of the SFP after interference, corresponding photoelectric conversion and analog-to-digital conversion processing are carried out in the SFP, and then digital signals are sent to an FPGA through an output end; the product structure is simplified on the whole, which is beneficial to improving the integration level, reducing the volume and reducing the equipment cost.

Preferably, the SFP comprises light emitting means, and light receiving and processing means;

the light emitting device adopts a DFB laser; the light receiving and processing device comprises a light measuring device, a transimpedance amplifier and a limiting amplifier which are used for sequentially processing signals, wherein the light measuring device is based on a PIN photodiode form or an APD photodiode form.

Preferably, a port output module connected with the FPGA is further arranged for outputting the quantum random number.

The port output module may take the form of, for example, a network port, a USB interface, an optical port, etc.

In terms of SFP and FPGA itself, the prior art may be adopted, and in order to adapt to the structural characteristics of SFP in the random number generator of the present invention, an improvement is made on an interference device, which is preferably a beam splitter, and the beam splitter includes two incident ends and two corresponding exit ends, where:

the first incident end is connected with the SFP emergent end through a first optical fiber;

the first emergent end is connected with the second incident end through a second optical fiber;

the second emergent end is connected with the SFP incident end, and optical signals from the first incident end and the second incident end are output to the SFP incident end through the second emergent end after interference;

a delay fiber is disposed in at least one of the first and second optical fibers.

Optionally, the first emitting end is a reflected light output end corresponding to the first incident end, and the second emitting end is a transmitted light output end corresponding to the first incident end.

In the scheme, optical signal transmission and interference between the SFP can be completed by only adopting one beam splitter, and the product structure is further simplified.

In the solution of the other interference device, preferably, the interference device includes two beam splitters, where an optical signal output by the SFP output end is split into two paths by the first beam splitter, one path is delayed and then both the delayed path and the other path are input into the second beam splitter and interfere with each other, and the interference is then transmitted to the SFP input end.

In a further embodiment of the interference device, the interference device preferably comprises a circulator, a beam splitter and two faraday rotator mirrors, wherein:

the first port of the circulator is connected with the SFP emergent end;

the optical signal output by the second port of the circulator is divided into two paths by a beam splitter, and the two paths enter a corresponding Faraday rotary mirror respectively, wherein at least one path is provided with a delay optical fiber;

the optical signals reflected by the Faraday rotary mirrors return to the beam splitter along the original optical path to be subjected to beam combination interference, and then are sequentially output to the SFP incidence end through the second port of the circulator and the third port of the circulator.

Preferably, the beam splitters in the interference device are polarization maintaining beam splitters, and the optical fibers used in the interference device are polarization maintaining fibers.

The invention has the technical effects that:

1. the DFB laser of the SFP is used as a light source, so that an external DFB laser is not needed, the structure is simplified, and the cost is saved.

And 2, the SFP has small volume, mature technology and contribution to further reducing the volume of the quantum random number generator and improving the system stability.

Drawings

FIG. 1 is a schematic diagram of a prior art quantum random number generator;

FIG. 2 is a schematic diagram of the structure of the quantum random number generator of the present invention;

FIG. 3 is a schematic diagram of the structure of an SFP;

FIG. 4 is a schematic diagram of the structure of the random number generator in embodiment 1;

FIG. 5 is a schematic diagram of the structure of a random number generator in embodiment 2;

fig. 6 is a schematic diagram of the structure of the random number generator in embodiment 3.

Detailed Description

Referring to fig. 2, the quantum random number generator in the embodiment of the present invention includes:

the SFP is provided with an emergent end for transmitting optical signals, an incident end for receiving the optical signals and an output end for outputting digital signals; it has the light source/detection function.

The interference device is used for receiving the optical signals of the SFP emergent end and transmitting the optical signals to the SFP incident end after interference;

and the FPGA is used for receiving the digital signal of the SFP output end and generating the quantum random number. The FPGA can be internally divided into a data acquisition module and a post-processing module;

the port output module is connected with the post-processing module, and the quantum random number generated by the post-processing module is input into subsequent random number using equipment through the port output module (such as a network port, a USB interface, an optical port and the like).

Referring to fig. 3, sfp is mainly composed of two major parts: light emitting means, and light receiving and processing means.

The light emitting device comprises a corresponding light source and a driving circuit; the light receiving and processing device mainly includes an optical measuring device (photoelectric conversion), a transimpedance amplifier (amplifying an electric signal), and a limiting amplifier (analog-to-digital conversion).

The light emitting device is preferably a DFB laser (distributed feedback laser), which is an ideal light source for interference; the optical measurement device is preferably based on the form of a PIN photodiode or APD photodiode (avalanche photodiode).

Example 1

Referring to fig. 4, the quantum random number generator in this embodiment further refines the interference device based on the principle of fig. 1, where the interference device uses a beam splitter with four ports, the beam splitter preferably uses a polarization maintaining beam splitter, and the required optical fibers in the scheme use polarization maintaining fibers.

The specific construction and connection of the interference device is further described in connection with the operation.

When the quantum random number generator works, the quantum random number generator comprises:

1. and adding a fixed level to the Tx port of the SFP, so that the DFB laser built in the SFP emits stable continuous laser through the emergent end of the DFB laser.

Dfb laser light enters the entrance end of the beam splitter (port 1).

3. The other incident end (port 2) of the beam splitter is connected with the corresponding reflecting end (port 3) of the port 1 by a fiber delay line.

4. The interfered laser exits through the port 4 of the beam splitter and then enters the incidence end of the SFP for photoelectric detection.

The measurement module built in the SFP converts the optical signal into an electrical signal, and then amplifies the electrical signal and performs single-bit analog-to-digital conversion.

And 6, enabling an output signal of the SFP to enter an FPGA, and digitally acquiring an input signal by a high-speed clock in the FPGA to obtain original random number data.

In order to determine the appropriate compression ratio, this step requires that a large amount of raw random number data be collected first for randomness analysis (see the effect test section below).

The original data can also be processed by adopting a multiple exclusive or processing mode.

7. And (3) compressing and extracting the original data by a Toeplitz-mapping program pre-written into the FPGA to obtain the final quantum random number.

8. The quantum random number is output through various ports (such as a network port, a USB interface, an optical port and the like), and then the corresponding random number using equipment is input.

Example 2

Referring to fig. 5, the quantum random number generator in this embodiment further refines the interference device based on the principle of fig. 1, where the interference device uses two beam splitters, each beam splitter preferably uses a polarization maintaining beam splitter, and the required optical fiber in the scheme uses a polarization maintaining optical fiber.

1. The laser light is split into two identical laser light beams via a first balanced beam splitter (BS 1).

2. The two laser beams respectively enter a second balance beam splitter (BS 2) for interference after being transmitted by the optical fibers. The two optical fibers have different lengths, so that two laser beams at different moments interfere. The two balanced splitters and the fiber form a common MZI interferometer.

3. The interfered laser enters the incidence end of the SFP for measurement.

The subsequent steps of this example are the same as those of example 1, and see steps 5 to 8 in example 1.

Example 3

Referring to fig. 6, the quantum random number generator in this embodiment further refines the interference device based on the principle of fig. 1, where the interference device uses a circulator, a beam splitter and two faraday rotation mirrors, the beam splitter preferably uses a polarization maintaining beam splitter, and the required optical fibers in the scheme use polarization maintaining optical fibers.

2. The laser enters port 1 of the three port fiber optic circulator and then exits port 2.

3. The laser light is split into two identical laser light beams via a first balanced Beam Splitter (BS).

4. The two laser beams are transmitted through optical fibers (one of which is delayed) respectively, and then reflected through corresponding first and second Faraday rotary mirrors (FM 1 and FM 2) respectively, and then returned to the BS through the optical fibers for interference. Since the lengths of the two optical fibers are different, the lasers at different times interfere. The use of a faraday rotator mirror can automatically compensate for polarization shifts that occur when laser light is transmitted in an optical fiber.

5. The interfered laser sequentially enters the incidence end of the SFP through the ports 2 and 3 of the three-port optical fiber circulator for measurement.

Quantum random number generator principle

The basic principle of the quantum random number generator based on laser phase randomness is spontaneous radiation. The electric field for a continuous laser can be described by the following formula

Where A is amplitude and ω is frequencyIs the fluctuation of phase. After a delay of a fixed tau time by a fiber delay line

Interfering the delayed and undelayed two laser beams, wherein the square strength after interference is

Here the constant term is ignored and the term is used,the fluctuation of the phase is converted into the fluctuation of the light intensity by interference, and the fluctuation of the light intensity can be simply measured by a broadcast television detector. True randomness results fromThis term contains both classical fluctuations due to classical noise and quantum fluctuations due to spontaneous radiation (quantum fluctuations will follow a gaussian distribution). By operating the DFB laser around the threshold value, the quantum fluctuation dominates the phase fluctuation, but at the same time, due to a certain correlation between random numbers generated by the influence of unavoidable classical fluctuation, conventional randomness tests cannot be passed, so that further post-processing is required to extract true random numbers.

The above-described interference formula is used to describe the interference schemes in embodiment 2 and embodiment 3, and for embodiment 1, the structure is relatively simple but the interference process is more complicated than the former, but the nature of randomness is the same as that of the above scheme, so that the description will not be repeated.

The randomness of the original data is evaluated by adopting the minimum entropy, and the calculation formula is as follows

n represents the value of the sample, i.e. the data is divided in units of n bits. Giving a minimum entropy rate according to a calculation result of the minimum entropy, wherein a calculation formula is l=H _∞ (X)/n, which indicates the upper limit of the quantum random bits that can be extracted from each original bit.

Effect testing

To demonstrate the reliability of the protocol, specific experiments were performed for each example and the data were analyzed accordingly. The analysis of the pattern is mainly performed by the data of the following three aspects: 1. minimum entropy of the original data; 2. minimum entropy of post-processed data; 3. and (5) counting the test results.

For example 1, the minimum entropy of the raw data is shown in the following table

n	Minimum entropy rate (raw data)	Minimum entropy rate (post-processing data)
			1	0.837680193	0.999968147
2	0.681031144	0.999936964
			3	0.664931504	0.999904251
4	0.630478646	0.999812051
			5	0.657811791	0.999834876
6	0.674639878	0.999745208
			7	0.684794855	0.999632498
8	0.655774098	0.999582662
			9	0.675624979	0.999216714
10	0.677261662	0.998911089

TABLE 1

The parameter conditions for the data acquisition are: the bandwidth of SFP is 1.25Gb/s, the sampling frequency of FPGA is 500Mbps, and the size of test data is 1.5Gb. For different numbers of sample bits (n), the minimum entropy data fluctuates, but is above 0.6 (in order to indicate the reasonability of the size of the test data, the test data is further improved to 8Gb, and the calculation result of the minimum entropy only has slight change). Therefore, in the subsequent post-processing process using Toeplitz-Hashing, 60% compression processing is performed on the original data (the output data is 60% of the original data), the processed data are shown in table.1, and the minimum entropy rates of different sample bit numbers are very close to 1, which indicates that the data obtained by the post-processing have removed classical correlations to obtain completely random quantum random numbers. The influence of the FPGA sampling frequency on the data is further analyzed in the experiment, after the sampling frequency is increased to 1.25GHz, the minimum entropy of the original data is obviously reduced, which indicates that the oversampling exists, so that the sampling rate of 500MHz is a reasonable choice. At the same time, the bandwidth analysis of the SFP shows that the SFP adopting the bandwidth of 2.5Gb/s has no obvious effect on the original data, so that the SFP adopting the bandwidth of 1.25Gb/s is a better choice of the scheme from the viewpoint of cost.

To further verify the randomness of the post-processed data, a current general random number statistical test procedure is used to perform a strict randomness test on the post-processed data. There are two test procedures used:

1. the random number test program provided by the national institute of standards, NIST test for short;

2.Alphabit Battery Test is currently the only randomness test software specifically designed for hardware random number generators.

The test results show that the data after post-treatment can pass the test smoothly, and the test results are shown in the following table

Subtest item	P value	Proportion of
			Frequency	0.406499	990/1000
Block Frequency	0.020548	984/1000
			Cumulative Sums*	0.419021	991/1000
Runs	0.304126	987/1000
			Longest Run	0.719747	992/1000
Rank	0.492436	986/1000
			FFT	0.861264	987/1000
Nonoverlapping Template*	0.000850	984/1000
			Overlapping Template	0.101311	990/1000
Universal	0.282626	989/1000
			Approximate Entropy	0.186566	988/1000
Random Excursions*	0.111669	604/615
			Random Excursions Variant*	0.036137	609/615
Serial*	0.371941	991/1000
			Linear Complexity	0.715679	990/1000

TABLE 2

Table 2 shows the specific results of NIST testing (test data size 1 Gb), with sub-test entries given multiple P values, only the worst set of values in the table. Two conditions need to be met to pass NIST testing: 1. p values for all subtest were not less than 0.0001;2. the passing ratio is not lower than a predetermined value. The 1000 groups of tests in Table 2 passed no less than 980 groups, while the 615 groups passed no less than 601 groups. The results of Table 2 demonstrate that post-processed data successfully passed NIST testing.

TABLE 3

Table.3 shows the test results of Alphabit Battery Test (test data size 1 Gb), with the sub-test entries given by x numbers giving multiple P values but only the worst of them in the table. To pass the test, the P value for each sub-test must be within the interval [0.001,0.999], so Table 3 shows that the post-processed data passes the test successfully.

Experimental test data for example 2 are given below, table 4 shows the minimum entropy comparison of the data before and after post-processing, all parameters being the same as above.

n	Minimum entropy rate (raw data)	Minimum entropy rate (post-processing data)
			1	0.994981456	0.999951123
2	0.746511952	0.999942123
			3	0.717614325	0.999898204
4	0.651577131	0.999865107
			5	0.673336656	0.9997986
6	0.652173162	0.999721117
			7	0.65620207	0.99950246
8	0.608201799	0.999241837
			9	0.648742092	0.998990979
10	0.640978523	0.998390052

TABLE 4

Table 5 shows the results of NIST test

TABLE 5

The test of Table 5, set 599, was passed by set 585, the remainder being the same as described above, and the test file size was also 1Gb. The data presented in Table 5 shows that post-processed data successfully passed NIST testing.

Table 6 shows the results of the alpha bit Battery test

Subtest item	P value
		Multinomial Bits Over(L＝2)*	0.99
Multinomial Bits Over(L＝4)*	0.06
		Multinomial Bits Over(L＝8)*	0.98
Multinomial Bits Over(L＝16)*	0.90
		Hamming Indep(L＝16)	0.68
Hamming Indep(L＝32)	0.13
		Hamming Corr	0.40
RandomWalk1 H(L＝64)	0.23
		RandomWalk1 M(L＝64)	0.47
RandomWalk1 J(L＝64)	0.58
		RandomWalk1 R(L＝64)	0.59
RandomWalk1 C(L＝64)	0.91
		RandomWalk1 H(L＝320)	0.32
RandomWalk1 M(L＝320)	0.25
		RandomWalk1 J(L＝320)	0.25
RandomWalk1 R(L＝320)	0.82
		RandomWalk1 C(L＝320)	0.04

TABLE 6

The data indicate that the post-processed data can pass the test successfully.

Since the scheme of embodiment 3 is substantially the same as that of embodiment 2, a detailed description thereof will be omitted.

The above disclosure is merely an example of the present invention, but the present invention is not limited thereto and those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present invention. It is apparent that such modifications and variations are intended to be within the scope of the invention as claimed. In addition, although specific terms are used in the present specification, these terms are for convenience of description only and do not constitute any particular limitation on the present invention.

Claims

1. A quantum random number generator based on SFP transceiving integration, comprising:

the interference device is used for receiving the optical signals at the SFP emergent end, transmitting the optical signals to the SFP incident end after interference, and converting the randomness of the laser phase into the randomness of the intensity; the interference device is a beam splitter, and the beam splitter comprises two incident ends and two corresponding emergent ends, wherein the first incident end is connected with the SFP emergent end through a first optical fiber; the first emergent end is connected with the second incident end through a second optical fiber; the second emergent end is connected with the SFP incident end, and optical signals from the first incident end and the second incident end are output to the SFP incident end through the second emergent end after interference; a delay fiber is configured on at least one of the first optical fiber and the second optical fiber;

the FPGA is used for carrying out digital acquisition and post-processing on the digital signal at the SFP output end to finally generate a quantum random number; the SFP comprises a light emitting device and a light receiving and processing device;

2. The SFP-transceiver-integrated quantum random number generator as recited in claim 1, further comprising a port output module connected to the FPGA for outputting quantum random numbers.

3. The SFP-transceiver-based quantum random number generator of claim 1, wherein the first output end is a reflected light output end corresponding to the first incident end, and the second output end is a transmitted light output end corresponding to the first incident end.

4. The quantum random number generator based on SFP transceiver as claimed in claim 1, wherein the interference device comprises two beam splitters, wherein the optical signal output from the SFP output end is split into two paths by the first beam splitter, one path is delayed and then is input to the second beam splitter with the other path to interfere, and the interference is transmitted to the SFP input end.

5. The SFP transceiver-based quantum random number generator of claim 1, wherein the interference device comprises a circulator, a beam splitter, and two faraday rotation mirrors, wherein:

the first port of the circulator is connected with the SFP emergent end;

6. The SFP-transceiver-based quantum random number generator as claimed in claim 1, 4 or 5, wherein the beam splitters in the interference devices are polarization maintaining beam splitters, and the optical fibers used in the interference devices are polarization maintaining fibers.