Password security measurement method based on large language model and multi-side channel discriminator

Abstract

The invention relates to the field of information security passwords, and discloses a password security measurement method based on a large language model and a multi-side channel differentiator, which comprises the following steps: analyzing the characteristics of SPA according to simple energy, and designing various side channel differentiators such as segmentation, preprocessing, data statistics, dimension reduction, clustering and continuous rule limitation based on common analysis means; when the password equipment runs a target algorithm, acquiring side information data such as illumination, energy or electromagnetism in the running process of the password equipment; based on the collected side information data, different side channel distinguishers are selected, expert version cues with side channel analysis knowledge are formed through fusion, and analysis targets and tasks are described. According to the method, security automatic measurement can be carried out on password equipment with multiple protection means such as random delay and mask by utilizing a general large language model without expert participation or professional analysis tools, and compared with a common SPA security measurement method, the false alarm rate gt is reduced under the condition of the same sample size; the sample size is reduced by gt under the same accuracy rate; and therefore, the accuracy of the side channel security measurement is greatly improved, and the cost of resources such as labor and the like is reduced.

Description

Technical Field

[0001] The present invention relates to the field of information security cryptography, and in particular to a cryptographic security measurement method based on a large language model and a multi-side channel distinguisher. Background Art

[0002] Side channel analysis is a powerful technical means that is widely used to evaluate the key information leakage of cryptographic devices. Among them, Simple Power Analysis (SPA) is one of the basic security measurement methods for side channel analysis. Although SPA has important technical value, its operation usually requires experts with professional knowledge and high-end analysis tools, which makes it difficult for non-professionals or amateur users to conduct relevant analysis effectively.

[0003] At the same time, the field of Large Language Model (LLM) has made breakthrough progress in recent years. For example, ChatGPT has attracted much attention for its significant efficiency improvement in natural language processing, task assistance and other fields. The capabilities of LLM have been widely used in many complex tasks, showing excellent assistance potential.

[0004] However, in the field of side channel analysis, especially in SPA tasks, how to utilize the powerful capabilities of LLM has not been studied and explored in depth. At present, the possibility of performing SPA tasks by sending a descriptive text and waveform data to LLM has not been studied and explored. The research and development of this method is expected to not only lower the technical threshold of SPA tasks, but also expand the application boundaries of LLM in the field of side channel security and provide new analysis tools and ideas. Summary of the invention

[0005] The purpose of the present invention is to creatively propose a cryptographic security measurement method based on a large language model and multiple side channel distinguishers in order to solve the technical problems such as the need for expert participation or high-end analysis tools in the side channel security measurement process in view of the shortcomings and defects of the prior art. This method does not require the use of specific side channel analysis tools or expert participation, and can use a universal large language model to automatically measure the security of cryptographic devices with multiple protection means such as random delays, masks, etc.

[0006] The present invention is implemented by adopting the following technical solutions.

[0007] The cryptographic security measurement method based on a large language model and a multi-side channel distinguisher includes the following steps: Step 1: According to the characteristics of simple energy analysis SPA, based on common analysis methods, design a variety of side channel distinguishers such as segmentation, preprocessing, data statistics, dimension reduction, clustering and continuous rule restriction; Step 2: When the cryptographic device runs the target algorithm, collects illumination, energy, or electromagnetic information data during its operation; Step 3: Based on the collected side information data, different side channel discriminators are selected and integrated to form expert prompt words with side channel analysis knowledge to describe the analysis goals and tasks; Step 4: Send the expert version of the prompt words and side information data to the large language model for security measurement. If a leak is found, it is marked as "failed"; if no leak is detected, it is marked as "passed".

[0008] According to an example of an embodiment of the present disclosure, the designed "segmentation" side channel distinguisher is "selecting appropriate methods such as equal spacing, finding peaks and troughs to segment the waveform", the designed "preprocessing" side channel distinguisher is "using sliding windows, filtering and other preprocessing methods to preprocess the waveform", the designed "data statistics" side channel distinguisher is "to perform statistics on the mean, variance, number of peaks, number of points and other data of the waveform", the designed "dimensionality reduction" side channel distinguisher is "considering the numerical data information of the waveform to select an appropriate method for dimensionality reduction", the designed "clustering" side channel distinguisher is "using clustering method to divide the waveform into two categories", and the designed "continuous rule restriction" side channel distinguisher is "divided into two types of operations, squaring and multiplication, the first operation must be a squaring operation and the squaring operation has a total of key bits, and there must be at least one squaring operation between two multiplication operations".

[0009] According to an example of an embodiment of the present disclosure, when the STM32F429 development board and the contact smart card run the RSA decryption algorithm, the energy data during the operation is collected, and the waveforms are as follows: Figure 2 , 3 shown.

[0010] According to an example of an embodiment of the present disclosure, based on the collected side information data, for the energy data collected by the STM32F429 development board, side channel discriminators such as segmentation, data statistics, clustering and continuous rule restrictions are selected to be fused to form an expert version of prompt words with side channel analysis knowledge; for the energy data collected by the contact smart card, side channel discriminators such as preprocessing, segmentation, dimensionality reduction, clustering and continuous rule restrictions are selected to be fused to form an expert version of prompt words with side channel analysis knowledge.

[0011] According to an example of an embodiment of the present disclosure, for the STM32F429 development board, the expert version prompt word and energy data fused in step 3 are sent to the large language model (ChatGPT-4), and its security measurement result is not passed. The result is as follows Figure 4As shown; for the contact smart card, the expert version prompt word and energy data fused in step 3 are sent to the large language model (ChatGPT-4), and its security measurement result is failed. The result is as follows Figure 5 shown.

[0012] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of the present application. Beneficial Effects

[0013] The present invention can effectively utilize commonly used analysis methods of side channels, without the need for expert participation or professional analysis tools, and can use a universal large language model to perform automated security measurement for cryptographic devices with multiple protection methods such as random delays and masks. Compared with the ordinary SPA security measurement method, the false alarm rate is reduced by >= 30% when the sample size is the same, and the sample size is reduced by >= 30% when the accuracy is the same, which greatly improves the accuracy of side channel security measurement and reduces resource costs such as labor. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The above and other purposes, features and advantages of the embodiments of the present disclosure will become more apparent by describing the embodiments of the present disclosure in more detail in conjunction with the accompanying drawings. The accompanying drawings are used to provide a further understanding of the embodiments of the present disclosure and constitute a part of the specification. Together with the embodiments of the present disclosure, they are used to explain the present disclosure and do not constitute a limitation of the present disclosure. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0015] Figure 1 is a flow chart of the method of the present invention; Figure 2 It is a schematic diagram of energy signals collected from the STM32F429 development board in the embodiment of the method of the present invention; Figure 3 is a schematic diagram of energy signals collected from a contact smart card in an embodiment of the method of the present invention; Figure 4 It is a schematic diagram of the security measurement results of the STM32F429 development board in the embodiment of the method of the present invention; Figure 5 Schematic diagram of security measurement results of a contact smart card in an embodiment of the method of the present invention. DETAILED DESCRIPTION

[0016] The technical solutions in the embodiments of the present disclosure will be described clearly and completely below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0017] In order to make the purpose, technical scheme and advantages of the present application clearer, the technical scheme of the present application is further elaborated in detail below in conjunction with the accompanying drawings and embodiments. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application. In the following description, it is related to "optional implementation method", which describes a subset of all possible embodiments, but it can be understood that the "optional implementation method" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict. The term "first / second" involved is only to distinguish similar objects, and does not represent a specific order for the object. It can be understood that "first / second" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described here can be implemented in an order other than that illustrated or described here. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of the present application. The terms used herein are only for the purpose of describing the present application and are not intended to limit the present application.

[0018] The embodiment of the present application provides, taking the RSA algorithm as an example, collecting the energy information of the RSA algorithm running on the STM32F429 development board and the contact smart card, and using the method of the present invention to perform side channel security measurement.

[0019] like Figure 1 As shown, the cryptographic security measurement method based on a large language model and a multi-side channel distinguisher includes the following steps: Step 1: According to the characteristics of simple energy analysis (SPA), based on common analysis methods, various side channel discriminators such as segmentation, preprocessing, data statistics, dimension reduction, clustering and continuous rule restriction are designed.

[0020] In the embodiment, the designed "segmentation" side channel distinguisher is "selecting appropriate methods such as equal spacing, finding peaks and troughs to segment the waveform", the designed "preprocessing" side channel distinguisher is "using sliding windows, filtering and other preprocessing methods to preprocess the waveform", the designed "data statistics" side channel distinguisher is "to perform statistics on the mean, variance, number of peaks, number of points and other data of the waveform", the designed "dimensionality reduction" side channel distinguisher is "considering the numerical data information of the waveform to select an appropriate method for dimensionality reduction", the designed "clustering" side channel distinguisher is "using clustering method to divide the waveform into two categories", and the designed "continuous rule restriction" side channel distinguisher is "dividing the operations into two categories of squaring and multiplication, the first operation must be a squaring operation and the squaring operation has a total of key bits, and there must be at least one squaring operation between two multiplication operations".

[0021] Step 2: When the cryptographic device runs the target algorithm, it collects illumination, energy, electromagnetic and other side information data during its operation.

[0022] In the embodiment, when the STM32F429 development board and the contact smart card run the RSA decryption algorithm, the energy data during the operation is collected, and the waveforms are as follows: Figure 2 , 3 shown.

[0023] Step 3: Based on the collected side information data, different side channel discriminators are selected and integrated to form expert prompts with side channel analysis knowledge to describe the analysis objectives and tasks.

[0024] In the embodiment, based on the collected side information data, for the energy data collected by the STM32F429 development board, side channel discriminators such as segmentation, data statistics, clustering and continuous rule restrictions are selected to be fused to form an expert version of prompt words with side channel analysis knowledge; for the energy data collected by the contact smart card, side channel discriminators such as preprocessing, segmentation, dimensionality reduction, clustering and continuous rule restrictions are selected to be fused to form an expert version of prompt words with side channel analysis knowledge.

[0025] Step 4: Send the expert version of the prompt words and side information data to the large language model for security measurement. If a leak is found, it is marked as "failed"; if no leak is detected, it is marked as "passed".

[0026] In the embodiment, for the STM32F429 development board, the expert version prompt word and energy data fused in step 3 are sent to the large language model (ChatGPT-4), and its security measurement result is not passed. The result is as follows Figure 4 As shown; for the contact smart card, the expert version prompt word and energy data fused in step 3 are sent to the large language model (ChatGPT-4), and its security measurement result is failed. The result is as follows Figure 5 shown.

[0027] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. The above-mentioned embodiments only express several implementation methods of the present application, and their descriptions are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of this application. It should be pointed out that for ordinary technicians in this field, without departing from the concept of the present application, several deformations and improvements can be made, which all belong to the scope of protection of the present application. Therefore, the scope of protection of the present application shall be based on the attached claims.

Claims

1. A cryptographic security measurement method based on a large language model and a multi-side channel distinguisher, characterized in that: The following steps are involved: Step 1: According to the characteristics of simple energy analysis SPA, based on common analysis methods, design a variety of side channel distinguishers such as segmentation, preprocessing, data statistics, dimension reduction, clustering and continuous rule restriction; Step 2: When the cryptographic device runs the target algorithm, collects illumination, energy, or electromagnetic information data during its operation; Step 3: Based on the collected side information data, different side channel discriminators are selected and integrated to form expert prompt words with side channel analysis knowledge to describe the analysis goals and tasks; Step 4: Send the expert version of the prompt words and side information data to the large language model for security measurement. If a leak is found, it is marked as "failed"; if no leak is detected, it is marked as "passed".

Images