WO2019136953A1 - 基于c&c域名分析的僵尸网络检测方法、装置、设备及介质 - Google Patents

基于c&c域名分析的僵尸网络检测方法、装置、设备及介质 Download PDF

Info

Publication number
WO2019136953A1
WO2019136953A1 PCT/CN2018/096107 CN2018096107W WO2019136953A1 WO 2019136953 A1 WO2019136953 A1 WO 2019136953A1 CN 2018096107 W CN2018096107 W CN 2018096107W WO 2019136953 A1 WO2019136953 A1 WO 2019136953A1
Authority
WO
WIPO (PCT)
Prior art keywords
domain name
botnet
domain
legal
category
Prior art date
Application number
PCT/CN2018/096107
Other languages
English (en)
French (fr)
Inventor
杜明
涂大志
王新成
Original Assignee
深圳市联软科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳市联软科技股份有限公司 filed Critical 深圳市联软科技股份有限公司
Publication of WO2019136953A1 publication Critical patent/WO2019136953A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L61/00Network arrangements, protocols or services for addressing or naming
    • H04L61/45Network directories; Name-to-address mapping
    • H04L61/4505Network directories; Name-to-address mapping using standardised directories; using standardised directory access protocols
    • H04L61/4511Network directories; Name-to-address mapping using standardised directories; using standardised directory access protocols using domain name system [DNS]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00Network architectures or network communication protocols for network security
    • H04L63/14Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
    • H04L63/1408Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
    • H04L63/1416Event detection, e.g. attack signature detection
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00Network architectures or network communication protocols for network security
    • H04L63/14Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
    • H04L63/1408Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
    • H04L63/1425Traffic logging, e.g. anomaly detection

Definitions

  • the present invention relates to the field of network security technologies, and in particular, to a botnet detection method, apparatus, device and medium based on C&C domain name analysis.
  • Botnet refers to the attacker or controller (Botmaster) propagating bots to control a large number of hosts, and through a network of one-to-many command and control channels, to achieve "send control commands to the controlled computer, indicating parasitic The purpose of the Trojan to perform a predetermined malicious action.
  • the controlled computer is called a broiler or zombie host or a bot for short, and Figure 1 is a botnet structure diagram.
  • botnet detection technologies mainly include: IDS (Instruction Detection System), honeypot technology and network traffic analysis.
  • IDS Intrusion Detection System
  • IDS Instruction Detection System
  • Honeypot technology seduce attacks by deliberately arranging the targets being attacked. Once an attacker invades, it can track how the attacks are implemented, analyze the connections between the attackers, and obtain their social networks.
  • honeypot technology requires a lot of deployment and is easily controlled as an attack springboard.
  • Network traffic analysis The research idea of network traffic is to analyze the behavior characteristics of zombie hosts in Botnet based on Internet Relay Chat (IRC) protocol, and classify zombie hosts into two categories: long-term stagnation and fast-joining. Specifically, there are three obvious behavioral characteristics of the zombie host in Botnet. One is a bot that spreads through the worm, and a large number of computers infected by it will join the same IRC Server in a short time; Second, the zombie host will generally be online for a long time; the third is that the zombie host acts as an IRC chat user, and does not speak for a long time in the chat channel, and remains idle. Traffic analysis can find some zombie hosts, but most of the malicious domain names randomly generated by the Command and Control server (C&C server) fail to generate traffic and random operation status of the network. It is difficult to lock the entire Internet in time and accurately. Zombie host, locate botnet.
  • C&C server Command and Control server
  • the existing botnet monitoring technology can not capture the attack behavior in time, lock the zombie host and locate the botnet.
  • the technical problem to be solved by the present application is to provide a botnet detection method, device, device and medium based on C&C domain name analysis, and analyze the domain name system (DNS) log record to extract the C&C domain name used by the attack activity, and then Analyze the type of parasitic Trojan, lock the zombie host controlled by the C&C server, and analyze the trend of botnet activity by analyzing the Poisson parameters generated by each type of C&C domain name to achieve timely and effective suppression measures.
  • DNS domain name system
  • the embodiment of the present application provides a botnet detection method based on C&C domain name analysis, including:
  • the domain name analysis step detects the C&C domain name in the DNS log record according to the pre-built domain name analyzer, and determines the category of each C&C domain name;
  • the botnet determination step determines whether a botnet exists based on the category of the C&C domain name and the C&C domain name.
  • the method further comprises:
  • Data statistics step counting the frequency of occurrence of each type of C&C domain name
  • the trend judgment step determines the activity trend of the botnet based on the frequency of occurrence of all categories of C&C domain names to assist in the timely development of effective suppression measures.
  • the trend determining step comprises:
  • the training process of the domain name analyzer comprises:
  • the domain name analyzer is trained according to the legal domain name training sample set, the C&C domain name training sample set, and the character probability dictionary.
  • the domain name analyzer is a neural network model based on a cumulative BP algorithm, and a regularization term that comprehensively considers an empirical error factor and a network complexity factor is provided in the neural network model.
  • the calculating step of the neural network model based on the cumulative BP algorithm comprises:
  • the stochastic gradient descent parameter is used to approximate the global minimum solution of the error function.
  • the domain name analysis step comprises:
  • the category of the C&C domain name is determined based on the classification number.
  • the embodiment of the present application provides a botnet detection apparatus based on C&C domain name analysis, including:
  • An information obtaining unit configured to acquire a DNS log record
  • a domain name analyzing unit configured to detect a C&C domain name in the DNS log record according to a pre-built domain name analyzer, and determine a category of each C&C domain name
  • the botnet determining unit is configured to determine whether a botnet exists according to the category of the C&C domain name and the C&C domain name.
  • the method further comprises:
  • the trend judging unit is configured to determine an activity trend of the botnet according to the frequency of occurrence of all categories of C&C domain names, so as to assist in formulating effective suppression measures in time.
  • the trend judging unit is configured to substitute the frequency of occurrence of each type of C&C domain name into a Poisson distribution probability function to obtain a Poisson parameter corresponding to the category; and determine all the Poisson parameters as a measure of botnet activity.
  • An indicator determining an activity trend of the botnet according to the metric of the botnet activity rule.
  • the training process of the domain name analyzer includes:
  • the domain name analyzer is a calculation of a neural network model based on a cumulative BP algorithm, and a regularization term that comprehensively considers an empirical error factor and a network complexity factor is provided in the neural network model.
  • the domain name analyzing unit is configured to:
  • an embodiment of the present application provides a computer device, including: at least one processor, at least one memory, and computer program instructions stored in a memory, which are implemented when the computer program instructions are executed by the processor. The method of the first aspect.
  • an embodiment of the present application provides a computer readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the method of the first aspect of the above embodiments.
  • the botnet detection method, device, device and medium based on C&C domain name analysis analyzes the domain name system (DNS) log record, extracts the C&C domain name used by the attack activity, and analyzes the type of the parasitic Trojan. Lock the zombie host controlled by the C&C server.
  • DNS domain name system
  • analyze the trend of botnet activity by analyzing the Poisson parameters generated by each type of C&C domain name, so as to implement effective suppression measures in time.
  • the frequency of occurrence of the C&C domain name and the Poisson parameter can be analyzed, and the trend of the botnet activity can be obtained, thereby facilitating the formulation of effective suppression measures.
  • FIG. 1 is a structural diagram of a botnet in the prior art provided by the present invention.
  • FIG. 2 is a flowchart of a botnet detection method based on C&C domain name analysis according to an embodiment of the present invention
  • FIG. 3 is still another flowchart of a botnet detection method based on C&C domain name analysis according to an embodiment of the present invention
  • FIG. 5 is a block diagram of a botnet detecting apparatus based on C&C domain name analysis according to an embodiment of the present invention
  • FIG. 6 is a schematic structural diagram of hardware of a computer device according to an embodiment of the present invention.
  • the botnet detection method based on C&C domain name analysis includes:
  • the information acquisition step S1 acquires a DNS log record
  • the domain name analysis step S2 detects the C&C domain name in the DNS log record according to the pre-built domain name analyzer, and determines the category of each C&C domain name;
  • the botnet determines step S3 to determine whether a botnet exists based on the C&C domain name and the category of the C&C domain name.
  • the botnet detection method based on C&C domain name analysis analyzes the domain name system (DNS) log record, extracts the C&C domain name used by the attack activity, analyzes the type of the parasitic Trojan, and locks the C&C server to control. Zombie host.
  • DNS domain name system
  • the format of the DNS log record is as shown in Table 1.
  • the domain name analysis is performed, and the domain name detection result as shown in Table 2 can be obtained, and in the domain name detection result, the C&C domain names belonging to the same category are counted in chronological order. .
  • the domain name analyzer in this embodiment can recognize 28 kinds of C&C domain names such as banjori.
  • the method further includes:
  • Data statistics step S4 counting the frequency of occurrence of each type of C&C domain name
  • the trend judging step S5 determines the activity trend of the botnet according to the frequency of occurrence of all categories of C&C domain names, so as to assist in formulating effective suppression measures in time.
  • the trend determining step S5 includes:
  • the zombie host requests a large number of new C&C domain names, most of which fail to resolve;
  • the C&C server domain name occurrence frequency satisfies the Poisson distribution.
  • the C&C domain name detection model judges the records extracted from the DNS logs, counts the number of times of occurrence of similar C&C domain name units, and substitutes the Poisson distribution probability function to estimate the Poisson parameter ⁇ for a certain period of time.
  • the Poisson distribution probability function is as follows:
  • the Poisson parameter is determined as a measure of botnet activity regularity
  • Table 3 is an analysis of the botnet activity trend.
  • any unit time can be used as the statistical period, and the average frequency is the number of C&C domain names captured in the current period of the period.
  • the zombie host IP address, MAC address
  • targeted suppression measures can be formulated in a timely manner.
  • the training process of the domain name analyzer includes:
  • the domain name analyzer is trained according to the legal domain name training sample set, the C&C domain name training sample set, and the character probability dictionary.
  • 1495163 legal domain names are published as valid domain names for websites published by Alexa and the like, and C&C domain names are obtained by using the public DGA (Domain Generated Algorithm) algorithm.
  • DGA Domain Generated Algorithm
  • Pseudo-random means that the string sequence seems to be random, but since its structure can be predetermined, it can be repeated and copied. This algorithm is often used in malware as well as remote control software.
  • the domain name characteristics are as shown in Table 4.
  • the domain name analyzer is a neural network model based on the cumulative BP algorithm, and the regularization term considering the empirical error factor and the network complexity factor is set in the neural network model.
  • the calculation steps of the neural network model based on the cumulative BP algorithm include: calculating the error objective function; describing the complexity of the neural network; estimating the model parameters by the cross-validation method; and using the stochastic gradient descent parameter to approximate the global minimum solution of the error function.
  • an n-gram (uni-gram, bi-gram, tri-gram) character probability dictionary is established by using 1495163 legal domain names obtained by cleaning.
  • the legal domain name is the same as all kinds of C&C domain names, and 1000 samples are randomly selected as the training sample set, and the cumulative BP algorithm is used.
  • the part describing the complexity of the neural network is added to the error objective function, and the model parameters are estimated by the cross-validation method.
  • the gradient descent adjusts the global minimum solution of the error function.
  • the regularization term is added in the BP algorithm training model process, and the empirical error and the network complexity are compromised, and the over-fitting can be effectively controlled.
  • the domain name analysis step S2 includes: extracting a domain name in the DNS log record; performing feature extraction on the domain name; determining whether the domain name is a C&C domain name according to the character probability dictionary; and performing domain name feature quantification on the C&C domain name To obtain the classification number of the C&C domain name; determine the category of the C&C domain name according to the classification number.
  • a botnet detection apparatus based on C&C domain name analysis provided by an embodiment of the present invention includes:
  • the information obtaining unit 1 is configured to acquire a DNS log record.
  • the domain name analyzing unit 2 is configured to detect a C&C domain name in the DNS log record according to a pre-built domain name analyzer, and determine a category of each C&C domain name;
  • the botnet determining unit 3 is configured to determine whether a botnet exists according to the category of the C&C domain name and the C&C domain name.
  • the botnet detection device based on the C&C domain name analysis provided by the embodiment of the present invention analyzes the domain name system (DNS) log record, extracts the C&C domain name used by the attack activity, analyzes the type of the parasitic Trojan, and locks the C&C server. Zombie host.
  • DNS domain name system
  • the format of the DNS log record is as shown in Table 1.
  • the domain name analysis is performed, and the domain name detection result as shown in Table 2 can be obtained, and in the domain name detection result, the C&C domain names belonging to the same category are counted in chronological order. .
  • the domain name analyzer in this embodiment can recognize 28 kinds of C&C domain names such as banjori.
  • the method further includes:
  • the data statistics unit 4 is configured to count the frequency of occurrence of each type of C&C domain name
  • the trend judging unit 5 is configured to determine the activity trend of the botnet according to the frequency of occurrence of all categories of C&C domain names, so as to assist in formulating effective suppression measures in time.
  • the trend judging unit 5 is specifically configured to:
  • the zombie host requests a large number of new C&C domain names, most of which fail to resolve;
  • the C&C server domain name occurrence frequency satisfies the Poisson distribution.
  • the C&C domain name detection model judges the records extracted from the DNS logs, counts the number of times of occurrence of similar C&C domain name units, and substitutes the Poisson distribution probability function to estimate the Poisson parameter ⁇ for a certain period of time.
  • the Poisson distribution probability function is as follows:
  • the Poisson parameter is determined as a measure of botnet activity regularity
  • Table 3 is an analysis of the botnet activity trend.
  • any unit time can be used as the statistical period, and the average frequency is the number of C&C domain names captured in the current period of the period.
  • the zombie host IP address, MAC address
  • targeted suppression measures can be formulated in a timely manner.
  • the training process of the domain name analyzer includes:
  • the domain name analyzer is trained according to the legal domain name training sample set, the C&C domain name training sample set, and the character probability dictionary.
  • 1,495,163 legal domain names are published as valid domain names for websites published by Alexa and the like, and C&C domain names are obtained by using the public DGA algorithm.
  • DGA is a domain name generation algorithm, and an attacker can use it to generate a pseudo-random string used as a domain name, so that the detection of the blacklist can be effectively avoided.
  • Pseudo-random means that the string sequence seems to be random, but since its structure can be predetermined, it can be repeated and copied. This algorithm is often used in malware as well as remote control software.
  • the domain name characteristics are as shown in Table 4.
  • the domain name analyzer is a calculation of a neural network model based on the cumulative BP algorithm, and the regularization term that comprehensively considers the empirical error factor and the network complexity factor is set in the neural network model.
  • the calculation steps of the neural network model based on the cumulative BP algorithm include: calculating the error objective function; describing the complexity of the neural network; estimating the model parameters by the cross-validation method; and using the stochastic gradient descent parameter to approximate the global minimum solution of the error function.
  • an n-gram (uni-gram, bi-gram, tri-gram) character probability dictionary is established by using 1495163 legal domain names obtained by cleaning.
  • the legal domain name is the same as all kinds of C&C domain names, and 1000 samples are randomly selected as the training sample set, and the cumulative BP algorithm is used.
  • the part describing the complexity of the neural network is added to the error objective function, and the model parameters are estimated by the cross-validation method.
  • the gradient descent adjusts the global minimum solution of the error function.
  • the feature is extracted according to the name character of the registration domain.
  • the regularization term is added to the BP algorithm training model to compromise the empirical error and network complexity, and the over-fitting can be effectively controlled.
  • the domain name analyzing unit 2 is specifically configured to: extract a domain name in the DNS log record; perform feature extraction on the domain name; determine whether the domain name is a C&C domain name according to the character probability dictionary; and perform a domain name on the C&C domain name. Feature quantification to obtain the classification number of the C&C domain name; the category of the C&C domain name is determined according to the classification number.
  • FIG. 6 is a schematic diagram showing the hardware structure of a computer device according to an embodiment of the present invention.
  • a computer device implementing a botnet detection method based on C&C domain name analysis may include a processor 401 and a memory 402 storing computer program instructions.
  • the processor 401 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present invention. .
  • CPU central processing unit
  • ASIC application specific integrated circuit
  • Memory 402 can include mass storage for data or instructions.
  • the memory 402 can include a Hard Disk Drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a Universal Serial Bus (USB) drive, or two or more. A combination of more than one of these.
  • Memory 402 may include removable or non-removable (or fixed) media, where appropriate.
  • Memory 402 may be internal or external to the data processing device, where appropriate.
  • memory 402 is a non-volatile solid state memory.
  • memory 402 includes a Read-Only Memory (ROM).
  • the ROM may be a mask-programmed ROM, a Programmable Read-only Memory (PROM), an Erasable Programmable ROM (EPROM), or an electrically erasable PROM (Electrically Erasable Programmable).
  • PROM Programmable Read-only Memory
  • EPROM Erasable Programmable ROM
  • PROM Electrically Erasable Programmable
  • EEPROM Electrically rewritable ROM
  • flash memory or a combination of two or more of these.
  • the processor 401 implements any of the above-described embodiments based on the C&C domain name analysis-based botnet detection method by reading and executing the computer program instructions stored in the memory 402.
  • the computer device can also include a communication interface 403 and a bus 410. As shown in FIG. 4, the processor 401, the memory 402, and the communication interface 403 are connected by the bus 410 and complete communication with each other.
  • the communication interface 403 is mainly used to implement communication between modules, devices, units and/or devices in the embodiments of the present invention.
  • Bus 410 includes hardware, software, or both that couples components of the computer device to each other.
  • the bus may include Accelerated Graphic Ports or Advanced Graphic Ports (AGP) or other graphics bus, Enhanced Industry Standard Architecture (EISA) bus, Front Side Bus (Front Side Bus, FSB), HyperTransport (HT) interconnect, Industry Standard Architecture (ISA) bus, infinite bandwidth interconnect, Low Pin Count (LPC) bus, memory bus, microchannel architecture ( MicroChannel Architecture, MCA) Bus, Peripheral Component Interconnect (PCI) bus, PCI-Express (PCI-X) bus, Serial Advanced Technology Attachment (SATA) bus, Video Electronics Standards Association (VESA local bus, VLB) bus or other suitable bus or a combination of two or more of these.
  • Bus 410 may include one or more buses, where appropriate. Although specific embodiments of the present invention are described and illustrated, the present invention contemplates any suitable bus or interconnect.
  • the embodiment of the present invention may be implemented by providing a computer readable storage medium.
  • the computer readable storage medium stores computer program instructions; when the computer program instructions are executed by the processor, the botnet detection method based on the C&C domain name analysis of any of the above embodiments is implemented.
  • the functional blocks shown in the block diagrams described above may be implemented as hardware, software, firmware, or a combination thereof.
  • hardware When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), suitable firmware, plug-ins, function cards, and the like.
  • ASIC application specific integrated circuit
  • the elements of the present invention are programs or code segments that are used to perform the required tasks.
  • the program or code segments can be stored in a machine readable medium or transmitted over a transmission medium or communication link through a data signal carried in the carrier.
  • a "machine-readable medium” can include any medium that can store or transfer information.
  • machine-readable media examples include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, and the like.
  • the code segments can be downloaded via a computer network such as the Internet, an intranet, and the like.
  • the exemplary embodiments referred to in the present invention describe some methods or systems based on a series of steps or devices.
  • the present invention is not limited to the order of the above steps, that is, the steps may be performed in the order mentioned in the embodiment, or may be different from the order in the embodiment, or several steps may be simultaneously performed.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Security & Cryptography (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer Hardware Design (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Debugging And Monitoring (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本发明提供一种基于C&C域名分析的僵尸网络检测方法、装置、设备及介质,方法包括:信息获取步骤,获取DNS日志记录;域名分析步骤,根据预先构建的域名分析器,检测DNS日志记录中的C&C域名,并判断每条C&C域名的所属类别;僵尸网络确定步骤,根据C&C域名及C&C域名的所属类别,确定是否存在僵尸网络。本发明提供的基于C&C域名分析的僵尸网络检测方法、装置、设备及介质,通过分析域名系统日志记录,提取攻击活动使用的C&C域名,进而分析寄生木马的类型,锁定C&C服务器已控制的僵尸主机,此外,利用分析每类C&C域名发生的泊松参数,分析僵尸网络活动的趋势,以实现及时制定有效的抑制措施。

Description

基于C&C域名分析的僵尸网络检测方法、装置、设备及介质 技术领域
本发明涉及网络安全技术领域,尤其涉及一种基于C&C域名分析的僵尸网络检测方法、装置、设备及介质。
背景技术
僵尸网络(Botnet),是指攻击者或控制者(Botmaster)传播僵尸程序控制大量主机,并通过一对多的命令与控制信道所组成的网络,实现“向被控制计算机发送控制指令,指示寄生木马执行预定恶意动作”的目的。其中,称被控制计算机为肉鸡或僵尸主机或简称bot机,且图1为僵尸网络结构图。
目前,僵尸网络检测技术主要包括:入侵检测系统(IDS,Instruction Detection System)、蜜罐技术和网络流量分析。
(1)入侵检测系统(IDS,Instruction Detection System)。IDS通过配置安全策略,对网络、系统的运行状况进行监视,尽可能发现各种攻击行为,以及时锁定感染主机,保证网络系统资源的机密性和可靠性。但是,IDS适合局域网环境,只能找到曾被发现的僵尸网络。
(2)蜜罐技术。蜜罐技术通过刻意布置被攻击的目标引诱攻击,一旦攻击者入侵后,就可以跟踪攻击如何实施、分析攻击者之间的相互联系,获取他们的社交网络。但是,蜜罐技术需要大量部署,且容易被控制作为攻击跳板。
(3)网络流量分析。网络流量的研究思路是通过分析基于互联网中继聊天(Internet Relay Chat,IRC)协议的Botnet中僵尸主机的行为特征,将僵尸主机分为两类:长时间发呆型和快速加入型。具体来说,僵尸主机在Botnet中存在着三个比较明显的行为特征,一是通过蠕虫传播的僵尸程序,大量的被其感染的计算机会在很短的时间内加入到同一个IRC Server中;二是僵尸主机一般 会长时间在线;三是僵尸主机作为一个IRC聊天的用户,在聊天频道内长时间不发言,保持空闲。流量分析可以找到部分僵尸主机,却因命令与控制服务器(Command and Control server,C&C服务器)随机生成的恶意域名大部解析失败不产生流量、网络的随机运行状态,很难及时准确地锁定整个互联网的僵尸主机、定位僵尸网络。
综上所述,现有的僵尸网络监测技术尚不能及时捕获攻击行为,锁定僵尸主机并定位僵尸网络。
发明内容
本申请要解决的技术问题是提供一种基于C&C域名分析的僵尸网络检测方法、装置、设备及介质,通过分析域名系统(Domain Name System,DNS)日志记录,提取攻击活动使用的C&C域名,进而分析寄生木马的类型,锁定C&C服务器已控制的僵尸主机,此外,利用分析每类C&C域名发生的泊松参数,分析僵尸网络活动的趋势,以实现及时制定有效的抑制措施。
第一方面,本申请实施例提供一种基于C&C域名分析的僵尸网络检测方法,包括:
信息获取步骤,获取DNS日志记录;
域名分析步骤,根据预先构建的域名分析器,检测DNS日志记录中的C&C域名,并判断每条C&C域名的所属类别;
僵尸网络确定步骤,根据C&C域名及C&C域名的所属类别,确定是否存在僵尸网络。
优选地,还包括:
数据统计步骤,统计每类C&C域名的发生频次;
趋势判断步骤,根据所有类别的C&C域名的发生频次,确定僵尸网络的活动趋势,以辅助及时制定有效的抑制措施。
优选地,趋势判断步骤,包括:
将每类C&C域名的发生频次代入泊松分布概率函数,以获取对应所属类别的泊松参数;
将所有泊松参数确定为僵尸网络活动规律衡量指标;
根据僵尸网络活动规律衡量指标,确定僵尸网络的活动趋势。
优选地,域名分析器的训练过程,包括:
对合法网站公开的合法域名进行清洗以获取合法域名集;
采用公开的域名生成算法生成C&C域名集,并对C&C域名集中的每个域名进行分类标记;
统计分析合法域名集和C&C域名集中的每个域名的结构,构建各类域名的字符概率字典;
从合法域名集中随机抽取设定数量的合法域名,获取合法域名训练样本集;
从C&C域名集中随机抽取设定数量的C&C域名,获取C&C域名训练样本集,
根据合法域名训练样本集、C&C域名训练样本集和字符概率字典,对域名分析器进行训练。
优选地,域名分析器为,基于累积BP算法的神经网络模型,且神经网络模型中设置有综合考虑经验误差因子和网络复杂度因子的正则化项。
优选地,基于累积BP算法的神经网络模型的计算步骤,包括:
计算误差目标函数;
描述神经网络复杂度;
通过交叉验证法估计模型参数;
使用随机梯度下降调参逼近误差函数全局最小解。
优选地,域名分析步骤,包括:
提取DNS日志记录中的域名;
对域名进行特征提取;
根据字符概率字典确定域名是否为C&C域名;
对C&C域名进行域名特征量化,以获取C&C域名的分类号;
根据分类号确定C&C域名的所属类别。
第二方面,本申请实施例提供一种基于C&C域名分析的僵尸网络检测装置,包括:
信息获取单元,用于获取DNS日志记录;
域名分析单元,用于根据预先构建的域名分析器,检测DNS日志记录中的C&C域名,并判断每条C&C域名的所属类别;
僵尸网络确定单元,用于根据C&C域名及C&C域名的所属类别,确定是否存在僵尸网络。
优选地,还包括:
数据统计单元,用于统计每类C&C域名的发生频次;
趋势判断单元,用于根据所有类别的C&C域名的发生频次,确定所述僵尸网络的活动趋势,以辅助及时制定有效的抑制措施。
优选地,所述趋势判断单元,用于将每类C&C域名的发生频次代入泊松分布概率函数,以获取对应所属类别的泊松参数;将所有所述泊松参数确定为僵尸网络活动规律衡量指标;根据所述僵尸网络活动规律衡量指标,确定所述僵尸网络的活动趋势。
优选地,所述域名分析器的训练过程,包括:
对合法网站公开的合法域名进行清洗以获取合法域名集;
采用域名生成算法生成C&C域名集,并对所述C&C域名集中的每个域名进行分类标记;
统计分析所述合法域名集和所述C&C域名集中的每个域名的结构,构建各类域名的字符概率字典;
从所述合法域名集中随机抽取设定数量的合法域名,获取合法域名训练样本集;
从所述C&C域名集中随机抽取设定数量的C&C域名,获取C&C域名训 练样本集,
根据所述合法域名训练样本集、所述C&C域名训练样本集和所述字符概率字典,对所述域名分析器进行训练。
优选地,所述域名分析器为,基于累积BP算法的神经网络模型的计算,且所述神经网络模型中设置有综合考虑经验误差因子和网络复杂度因子的正则化项。
优选地,所述域名分析单元用于:
提取DNS日志记录中的域名;
对所述域名进行特征提取;
根据所述字符概率字典确定所述域名是否为C&C域名;
对所述C&C域名进行域名特征量化,以获取所述C&C域名的分类号;
根据所述分类号确定所述C&C域名的所属类别。
第三方面,本申请实施例提供了一种计算机设备,包括:至少一个处理器、至少一个存储器以及存储在存储器中的计算机程序指令,当计算机程序指令被处理器执行时实现如上述实施方式中第一方面的方法。
第四方面,本申请实施例提供了一种计算机可读存储介质,其上存储有计算机程序指令,当计算机程序指令被处理器执行时实现如上述实施方式中第一方面的方法。
本申请实施例提供的基于C&C域名分析的僵尸网络检测方法、装置、设备及介质,通过分析域名系统(Domain Name System,DNS)日志记录,提取攻击活动使用的C&C域名,进而分析寄生木马的类型,锁定C&C服务器已控制的僵尸主机,此外,利用分析每类C&C域名发生的泊松参数,分析僵尸网络活动的趋势,以实现及时制定有效的抑制措施。
本申请实施例的有益技术效果是:
1、能够有效避免C&C域名绕过黑名单检测。
2、能够在网络中C&C攻击发起后,域名解析失败而未产生攻击流量时,及时捕获攻击行为。
3、能够在C&C服务器控制部分僵尸主机的情况下,分析C&C域名的发生频数及泊松参数,可获得僵尸网络活动趋势,从而有利于制定有效的抑制措施。
附图说明
图1是本发明提供的现有技术中的僵尸网络结构图;
图2是本发明实施例提供的基于C&C域名分析的僵尸网络检测方法的流程图;
图3是本发明实施例提供的基于C&C域名分析的僵尸网络检测方法的又一流程图;
图4是本发明实施例提供的C&C域名分类流程图;
图5是本发明实施例提供的基于C&C域名分析的僵尸网络检测装置的框图;
图6是本发明实施例提供的计算机设备的硬件结构示意图。
具体实施方式
下面通过具体的实施例进一步说明本发明,但是,应当理解为,这些实施例仅仅是用于更详细具体地说明之用,而不应理解为用于以任何形式限制本发明。
实施例一
结合图2,本实施例提供的基于C&C域名分析的僵尸网络检测方法,包括:
信息获取步骤S1,获取DNS日志记录;
域名分析步骤S2,根据预先构建的域名分析器,检测DNS日志记录中的C&C域名,并判断每条C&C域名的所属类别;
僵尸网络确定步骤S3,根据C&C域名及C&C域名的所属类别,确定是否存在僵尸网络。
本发明实施例提供的基于C&C域名分析的僵尸网络检测方法,通过分析域名系统(Domain Name System,DNS)日志记录,提取攻击活动使用的C&C域名,进而分析寄生木马的类型,锁定C&C服务器已控制的僵尸主机。本实施例中,具体地,DNS日志记录的格式如表1所示。
表1 DNS日志记录
时间 设备IP地址 域名 回应IP地址 TTL
2017-12-12 08:12:15.386 192.168.2.14 mbd.baidu.com 14.251.177.166 55
2017-12-12 08:12:15.889 192.168.2.19 news.ifeng.com 125.90.47.177 55
2017-12-12 08:12:16.231 192.168.2.110 www.78.cn 183.6.224.102 55
2017-12-12 08:12:17.001 192.168.2.118 www.ggspyfmreouxnhqi.com Null 0
2017-12-12 08:12:17.653 192.168.2.118 www.wyuhdsdttczd.com Null 0
2017-12-12 08:12:17.967 192.168.2.118 mail.pivzovznpssx.com Null 0
2017-12-12 08:12:18.862 192.168.2.118 www.swtjyuhuefvl.com Null 0
2017-12-12 08:12:19.768 192.168.2.118 www.zrkdvzjhse.com Null 0
2017-12-12 08:12:20.662 192.168.2.118 www.wyuhdsdttczd.com Null 0
2017-12-12 08:12:21.524 192.168.2.19 www.rauggyguyp.com 208.100.26.251 235
2017-12-12 08:12:22.325 192.168.2.118 www.furiararji.com Null 0
2017-12-12 08:12:23.219 192.168.2.118 www.pibqzedhzwt.com Null 0
2017-12-12 08:12:24.165 192.168.2.118 www.xjjcditjfkgkihfe.com Null 0
2017-12-12 08:12:24.981 192.168.2.14 tech.meituan.com 103.37.152.63 41
2017-12-12 08:12:25.824 192.168.2.19 www.iteblog.com 123.206.77.132 53
2017-12-12 08:12:26.585 192.168.2.110 guanjia.qq.com 14.215.138.13 55
2017-12-12 08:12:27.186 192.168.2.118 en.wikipedia.org 198.35.26.96 51
2017-12-12 08:12:28.115 192.168.2.118 www.johannesbader.ch 162.254.250.112 44
2017-12-12 08:12:29.023 192.168.2.14 us.norton.com 23.193.116.250 53
2017-12-12 08:12:29.829 192.168.2.118 www.swtjyuhuefvl.com Null 0
2017-12-12 08:12:30.691 192.168.2.110 spark.apache.org 195.154.151.36 50
2017-12-12 08:12:31.551 192.168.2.110 www.cnblogs.com 101.37.113.127 40
2017-12-12 08:12:32.384 192.168.2.14 blog.csdn.net 47.95.165.112 35
2017-12-12 08:12:33.168 192.168.2.19 baike.baidu.com 180.149.131.247 54
2017-12-12 08:12:34.069 192.168.2.118 www.jsntwyjcv.com Null 0
2017-12-12 08:12:35.011 192.168.2.118 app.tanwan.com 113.96.154.108 55
2017-12-12 08:12:35.892 192.168.2.110 www.icbc.com.cn 14.119.125.23 55
2017-12-12 08:12:36.721 192.168.2.118 www.miercn.com 113.96.154.108 55
2017-12-12 08:12:37.259 192.168.2.14 zs.91.com 125.77.24.228 53
2017-12-12 08:12:38.172 192.168.2.118 www.xjjcditjfkgkihfe.com Null 0
且本实施例中,针对如表1所示的日志记录,进行域名分析,能够获得如表2所示的域名检测结果,且域名检测结果中,按照时间顺序将属于同一类别的C&C域名统计出。
表2域名检测结果
时间 设备IP地址 域名 回应IP地址 TTL 类别
2017-12-12 08:12:17.001 192.168.2.118 www.ggspyfmreouxnhqi.com null 0 banjori
2017-12-12 08:12:17.653 192.168.2.118 www.wyuhdsdttczd.com null 0 banjori
2017-12-12 08:12:17.967 192.168.2.118 mail.pivzovznpssx.com null 0 banjori
2017-12-12 08:12:18.862 192.168.2.118 www.swtjyuhuefvl.com null 0 banjori
2017-12-12 08:12:19.768 192.168.2.118 www.zrkdvzjhse.com null 0 banjori
2017-12-12 08:12:21.524 192.168.2.19 www.rauggyguyp.com 208.100.26.251 235 banjori
2017-12-12 08:12:20.662 192.168.2.118 www.wyuhdsdttczd.com null 0 banjori
2017-12-12 08:12:22.325 192.168.2.118 www.furiararji.com null 0 banjori
2017-12-12 08:12:23.219 192.168.2.118 www.pibqzedhzwt.com null 0 banjori
2017-12-12 08:12:24.165 192.168.2.118 www.xjjcditjfkgkihfe.com null 0 banjori
2017-12-12 08:12:29.829 192.168.2.118 www.swtjyuhuefvl.com null 0 banjori
2017-12-12 08:12:34.069 192.168.2.118 www.jsntwyjcv.com null 0 banjori
2017-12-12 08:12:38.172 192.168.2.118 www.xjjcditjfkgkihfe.com null 0 banjori
此外,需要说明的是,本实施例中的域名分析器,能够对banjori等28种C&C域名进行识别。
优选地,如图3所示地,还包括:
数据统计步骤S4,统计每类C&C域名的发生频次;
趋势判断步骤S5,根据所有类别的C&C域名的发生频次,确定僵尸网络的活动趋势,以辅助及时制定有效的抑制措施。
且具体地,趋势判断步骤S5,包括:
将每类C&C域名的发生频次代入泊松分布概率函数,以获取对应所属类别的泊松参数;
将所有泊松参数确定为僵尸网络活动规律衡量指标;
根据僵尸网络活动规律衡量指标,确定僵尸网络的活动趋势。
本实施例中,出于经济成本,僵尸网络控制者不可能注册全部生成域名,仅事先注册若干生成域名。对于僵尸主机,为实现与C&C服务器建立连接,每个周期必生成同类的C&C域名尝试请求,直至获取C&C服务器的IP地址。于是,与正常主机相比,其行为模式有显著特征,主要表现为:
(1)僵尸主机请求大量新C&C域名,其中多数解析失败;
(2)当网络中存在多个寄生木马时,僵尸主机在域名请求行为呈现出行为特征,而且僵尸网络控制者拥有的服务器资源有限,其解析成功的C&C域名往往指向相同IP地址。
根据随机服务系统原理,C&C服务器域名发生频数满足泊松分布。由C&C域名检测模型判断从DNS日志提取的记录,统计同类C&C域名单位时间发生次数k,并代入泊松分布概率函数以估算某时段泊松参数λ,其中,泊松分布概率函数如下:
Figure PCTCN2018096107-appb-000001
本实施例中,将泊松参数确定为僵尸网络活动规律衡量指标,且表3为分析所得的僵尸网络活动趋势。
表3僵尸网络活动趋势
时段 平均频数 泊松参数 类别
01 45 45 Banjori
01 87 87 Sisron
01 0 0 Qadars
02 12 12 Banjori
02 0 0 Sisron
02 0 0 Qadars
03 53 53 Banjori
03 89 89 Sisron
03 36 36 Qadars
…… …… …… ……
表3中,任意单位时间均可作为统计时段,平均频数为周期内当前时段捕获C&C域名个数。
此外,需要说明的是,根据DNS日志记录,确定呈规律性发生的C&C域 名请求行为的僵尸主机(IP地址、MAC地址),至此根据发现僵尸主机的作用,容易分析该僵尸网络可能的攻击目标,可以及时制定针对性的抑制措施。
优选地,如图4所示地,域名分析器的训练过程,包括:
对合法网站公开的合法域名进行清洗以获取合法域名集;
采用公开的域名生成算法生成C&C域名集,并对C&C域名集中的每个域名进行分类标记;
统计分析合法域名集和C&C域名集中的每个域名的结构,构建各类域名的字符概率字典;
从合法域名集中随机抽取设定数量的合法域名,获取合法域名训练样本集;
从C&C域名集中随机抽取设定数量的C&C域名,获取C&C域名训练样本集,
根据合法域名训练样本集、C&C域名训练样本集和字符概率字典,对域名分析器进行训练。
本实施例中,对Alexa等网站公布的合法域名清洗获得1495163条作为合法域名,C&C域名均采用公开的DGA(Domain GenerateAlgorithm)算法采样获得。需要说明的是,DGA为域名生成算法,攻击者可以利用它来生成用作域名的伪随机字符串,这样就可以有效的避开黑名单列表的检测。伪随机意味着字符串序列似乎是随机的,但由于其结构可以预先确定,因此可以重复产生和复制。该算法常被运用于恶意软件以及远程控制软件上。本实施例中,域名特征如表4所示。
表4域名特征说明
特征名称 特征说明
length 主机名字符串长度
uni-entropy 主机名1-gram字符信息熵
uni-probavg 主机名1-gram字符平均概率
bi-entropy 主机名2-gram字符信息熵
bi-probavg 主机名2-gram字符平均概率
tri-entropy 主机名3-gram字符信息熵
tri-probavg 主机名3-gram字符平均概率
uni-gram-avgrank 主机名1-gram字符平均序
uni-gram-stdrank 主机名1-gram字符序标准差
bi-gram-avgrank 主机名2-gram字符平均序
bi-gram-stdrank 主机名2-gram字符序标准差
tri-gram-avgrank 主机名3-gram字符平均序
tri-gram-stdrank 主机名3-gram字符序标准差
vowel-ratio 元音字母占比
digit-ratio 数字占比
consonant-ratio 辅音字母占比
consec-consonant 连续辅音字母比例
consec-digit 连续数字比例
top1gram-ratio 主机名中1-gram字母概率top10比例
top2gram-ratio 主机名中2-gram字符组合概率top100比例
top3gram-ratio 主机名中3-gram字符组合概率top1000比例
本实施例中,具体地,域名分析器为,基于累积BP算法的神经网络模型的,且神经网络模型中设置有综合考虑经验误差因子和网络复杂度因子的正则化项。此外,基于累积BP算法的神经网络模型的计算步骤,包括:计算误差目标函数;描述神经网络复杂度;通过交叉验证法估计模型参数;使用随机梯度下降调参逼近误差函数全局最小解。本实施例中,利用清洗获得的1495163条合法域名,建立n-gram(uni-gram、bi-gram、tri-gram)字符概率字典。此外,合法域名与各类C&C域名一样,随机抽取1000条作为训练样本集,采用累积BP算法,并在误差目标函数中加入描述神经网络复杂度的部分,通过交叉验证法估计模型参数,使用随机梯度下降调参逼近误差函数全局最小解。
需要说明的是,本实施例根据注册域名字符习惯提取特征,BP算法训练模型过程中加入正则化项,对经验误差与网络复杂度进行折中,能够有效控制过拟合。
进一步优选地,如图4所示地,域名分析步骤S2,包括:提取DNS日志记录中的域名;对域名进行特征提取;根据字符概率字典确定域名是否为C&C域名;对C&C域名进行域名特征量化,以获取C&C域名的分类号;根据分类号确定C&C域名的所属类别。
实施例二
结合图5,本发明实施例提供的基于C&C域名分析的僵尸网络检测装置,包括:
信息获取单元1,用于获取DNS日志记录;
域名分析单元2,用于根据预先构建的域名分析器,检测DNS日志记录中的C&C域名,并判断每条C&C域名的所属类别;
僵尸网络确定单元3,用于根据C&C域名及C&C域名的所属类别,确定是否存在僵尸网络。
本发明实施例提供的基于C&C域名分析的僵尸网络检测装置,通过分析域名系统(Domain Name System,DNS)日志记录,提取攻击活动使用的C&C域名,进而分析寄生木马的类型,锁定C&C服务器已控制的僵尸主机。本实施例中,具体地,DNS日志记录的格式如表1所示。
表1 DNS日志记录
时间 设备IP地址 域名 回应IP地址 TTL
2017-12-12 08:12:15.386 192.168.2.14 mbd.baidu.com 14.251.177.166 55
2017-12-12 08:12:15.889 192.168.2.19 news.ifeng.com 125.90.47.177 55
2017-12-12 08:12:16.231 192.168.2.110 www.78.cn 183.6.224.102 55
2017-12-12 08:12:17.001 192.168.2.118 www.ggspyfmreouxnhqi.com Null 0
2017-12-12 08:12:17.653 192.168.2.118 www.wyuhdsdttczd.com Null 0
2017-12-12 08:12:17.967 192.168.2.118 mail.pivzovznpssx.com Null 0
2017-12-12 08:12:18.862 192.168.2.118 www.swtjyuhuefvl.com Null 0
2017-12-12 08:12:19.768 192.168.2.118 www.zrkdvzjhse.com Null 0
2017-12-12 08:12:20.662 192.168.2.118 www.wyuhdsdttczd.com Null 0
2017-12-12 08:12:21.524 192.168.2.19 www.rauggyguyp.com 208.100.26.251 235
2017-12-12 08:12:22.325 192.168.2.118 www.furiararji.com Null 0
2017-12-12 08:12:23.219 192.168.2.118 www.pibqzedhzwt.com Null 0
2017-12-12 08:12:24.165 192.168.2.118 www.xjjcditjfkgkihfe.com Null 0
2017-12-12 08:12:24.981 192.168.2.14 tech.meituan.com 103.37.152.63 41
2017-12-12 08:12:25.824 192.168.2.19 www.iteblog.com 123.206.77.132 53
2017-12-12 08:12:26.585 192.168.2.110 guanjia.qq.com 14.215.138.13 55
2017-12-12 08:12:27.186 192.168.2.118 en.wikipedia.org 198.35.26.96 51
2017-12-12 08:12:28.115 192.168.2.118 www.johannesbader.ch 162.254.250.112 44
2017-12-12 08:12:29.023 192.168.2.14 us.norton.com 23.193.116.250 53
2017-12-12 08:12:29.829 192.168.2.118 www.swtjyuhuefvl.com Null 0
2017-12-12 08:12:30.691 192.168.2.110 spark.apache.org 195.154.151.36 50
2017-12-12 08:12:31.551 192.168.2.110 www.cnblogs.com 101.37.113.127 40
2017-12-12 08:12:32.384 192.168.2.14 blog.csdn.net 47.95.165.112 35
2017-12-12 08:12:33.168 192.168.2.19 baike.baidu.com 180.149.131.247 54
2017-12-12 08:12:34.069 192.168.2.118 www.jsntwyjcv.com Null 0
2017-12-12 08:12:35.011 192.168.2.118 app.tanwan.com 113.96.154.108 55
2017-12-12 08:12:35.892 192.168.2.110 www.icbc.com.cn 14.119.125.23 55
2017-12-12 08:12:36.721 192.168.2.118 www.miercn.com 113.96.154.108 55
2017-12-12 08:12:37.259 192.168.2.14 zs.91.com 125.77.24.228 53
2017-12-12 08:12:38.172 192.168.2.118 www.xjjcditjfkgkihfe.com Null 0
且本实施例中,针对如表1所示的日志记录,进行域名分析,能够获得如表2所示的域名检测结果,且域名检测结果中,按照时间顺序将属于同一类别的C&C域名统计出。
表2域名检测结果
时间 设备IP地址 域名 回应IP地址 TTL 类别
2017-12-12 08:12:17.001 192.168.2.118 www.ggspyfmreouxnhqi.com null 0 banjori
2017-12-12 08:12:17.653 192.168.2.118 www.wyuhdsdttczd.com null 0 banjori
2017-12-12 08:12:17.967 192.168.2.118 mail.pivzovznpssx.com null 0 banjori
2017-12-12 08:12:18.862 192.168.2.118 www.swtjyuhuefvl.com null 0 banjori
2017-12-12 08:12:19.768 192.168.2.118 www.zrkdvzjhse.com null 0 banjori
2017-12-12 08:12:21.524 192.168.2.19 www.rauggyguyp.com 208.100.26.251 235 banjori
2017-12-12 08:12:20.662 192.168.2.118 www.wyuhdsdttczd.com null 0 banjori
2017-12-12 08:12:22.325 192.168.2.118 www.furiararji.com null 0 banjori
2017-12-12 08:12:23.219 192.168.2.118 www.pibqzedhzwt.com null 0 banjori
2017-12-12 08:12:24.165 192.168.2.118 www.xjjcditjfkgkihfe.com null 0 banjori
2017-12-12 08:12:29.829 192.168.2.118 www.swtjyuhuefvl.com null 0 banjori
2017-12-12 08:12:34.069 192.168.2.118 www.jsntwyjcv.com null 0 banjori
2017-12-12 08:12:38.172 192.168.2.118 www.xjjcditjfkgkihfe.com null 0 banjori
此外,需要说明的是,本实施例中的域名分析器,能够对banjori等28种C&C域名进行识别。
优选地,如图3所示地,还包括:
数据统计单元4,用于统计每类C&C域名的发生频次;
趋势判断单元5,用于根据所有类别的C&C域名的发生频次,确定僵尸网 络的活动趋势,以辅助及时制定有效的抑制措施。
且具体地,趋势判断单元5,具体用于:
将每类C&C域名的发生频次代入泊松分布概率函数,以获取对应所属类别的泊松参数;
将所有泊松参数确定为僵尸网络活动规律衡量指标;
根据僵尸网络活动规律衡量指标,确定僵尸网络的活动趋势。
本实施例中,出于经济成本,僵尸网络控制者不可能注册全部生成域名,仅事先注册若干生成域名。对于僵尸主机,为实现与C&C服务器建立连接,每个周期必生成同类的C&C域名尝试请求,直至获取C&C服务器的IP地址。于是,与正常主机相比,其行为模式有显著特征,主要表现为:
(1)僵尸主机请求大量新C&C域名,其中多数解析失败;
(2)当网络中存在多个寄生木马时,僵尸主机在域名请求行为呈现组行为特征,而且僵尸网络控制者拥有的服务器资源有限,其解析成功的C&C域名往往指向相同IP地址。
根据随机服务系统原理,C&C服务器域名发生频数满足泊松分布。由C&C域名检测模型判断从DNS日志提取的记录,统计同类C&C域名单位时间发生次数k,并代入泊松分布概率函数以估算某时段泊松参数λ,其中,泊松分布概率函数如下:
Figure PCTCN2018096107-appb-000002
本实施例中,将泊松参数确定为僵尸网络活动规律衡量指标,且表3为分析所得的僵尸网络活动趋势。
表3僵尸网络活动趋势
时段 平均频数 泊松参数 类别
01 45 45 Banjori
01 87 87 Sisron
01 0 0 Qadars
02 12 12 Banjori
02 0 0 Sisron
02 0 0 Qadars
03 53 53 Banjori
03 89 89 Sisron
03 36 36 Qadars
…… …… …… ……
表3中,任意单位时间均可作为统计时段,平均频数为周期内当前时段捕获C&C域名个数。
此外,需要说明的是,根据DNS日志记录,确定呈规律性发生的C&C域名请求行为的僵尸主机(IP地址、MAC地址),至此根据发现僵尸主机的作用,容易分析该僵尸网络可能的攻击目标,可以及时制定针对性的抑制措施。
优选地,如图4所示地,域名分析器的训练过程,包括:
对合法网站公开的合法域名进行清洗以获取合法域名集;
采用公开的域名生成算法生成C&C域名集,并对C&C域名集中的每个域名进行分类标记;
统计分析合法域名集和C&C域名集中的每个域名的结构,构建各类域名的字符概率字典;
从合法域名集中随机抽取设定数量的合法域名,获取合法域名训练样本集;
从C&C域名集中随机抽取设定数量的C&C域名,获取C&C域名训练样本集,
根据合法域名训练样本集、C&C域名训练样本集和字符概率字典,对域名分析器进行训练。
本实施例中,对Alexa等网站公布的合法域名清洗获得1495163条作为合法域名,C&C域名均采用公开的DGA算法采样获得。需要说明的是,DGA为域名生成算法,攻击者可以利用它来生成用作域名的伪随机字符串,这样就可以有效的避开黑名单列表的检测。伪随机意味着字符串序列似乎是随机的,但由于其结构可以预先确定,因此可以重复产生和复制。该算法常被运用于恶意软件以及远程控制软件上。本实施例中,域名特征如表4所示。
表4域名特征说明
特征名称 特征说明
length 主机名字符串长度
uni-entropy 主机名1-gram字符信息熵
uni-probavg 主机名1-gram字符平均概率
bi-entropy 主机名2-gram字符信息熵
bi-probavg 主机名2-gram字符平均概率
tri-entropy 主机名3-gram字符信息熵
tri-probavg 主机名3-gram字符平均概率
uni-gram-avgrank 主机名1-gram字符平均序
uni-gram-stdrank 主机名1-gram字符序标准差
bi-gram-avgrank 主机名2-gram字符平均序
bi-gram-stdrank 主机名2-gram字符序标准差
tri-gram-avgrank 主机名3-gram字符平均序
tri-gram-stdrank 主机名3-gram字符序标准差
vowel-ratio 元音字母占比
digit-ratio 数字占比
consonant-ratio 辅音字母占比
consec-consonant 连续辅音字母比例
consec-digit 连续数字比例
top1gram-ratio 主机名中1-gram字母概率top10比例
top2gram-ratio 主机名中2-gram字符组合概率top100比例
top3gram-ratio 主机名中3-gram字符组合概率top1000比例
本实施例中,具体地,域名分析器为,基于累积BP算法的神经网络模型的计算,且神经网络模型中设置有综合考虑经验误差因子和网络复杂度因子的正则化项。此外,基于累积BP算法的神经网络模型的计算步骤,包括:计算误差目标函数;描述神经网络复杂度;通过交叉验证法估计模型参数;使用随机梯度下降调参逼近误差函数全局最小解。本实施例中,利用清洗获得的1495163条合法域名,建立n-gram(uni-gram、bi-gram、tri-gram)字符概率字典。此外,合法域名与各类C&C域名一样,随机抽取1000条作为训练样本集,采用累积BP算法,并在误差目标函数中加入描述神经网络复杂度的部分,通过交叉验证法估计模型参数,使用随机梯度下降调参逼近误差函数全局最小解。
需要说明的是,本实施例根据注册域的名字符习惯提取特征。在BP算法训练模型中加入正则化项,以对经验误差与网络复杂度进行折中考虑,能够有 效控制过拟合。
进一步优选地,如图4所示地,域名分析单元2,具体用于:提取DNS日志记录中的域名;对域名进行特征提取;根据字符概率字典确定域名是否为C&C域名;对C&C域名进行域名特征量化,以获取C&C域名的分类号;根据分类号确定C&C域名的所属类别。
实施例三
结合图6描述的本发明实施例的基于C&C域名分析的僵尸网络检测方法可以由计算机设备来实现。图6示出了本发明实施例提供的计算机设备的硬件结构示意图。
实现基于C&C域名分析的僵尸网络检测方法的计算机设备可以包括处理器401以及存储有计算机程序指令的存储器402。
具体地,上述处理器401可以包括中央处理器(Central Processing Unit,CPU),或者特定集成电路(Application Specific Integrated Circuit,ASIC),或者可以被配置成实施本发明实施例的一个或多个集成电路。
存储器402可以包括用于数据或指令的大容量存储器。举例来说而非限制,存储器402可包括硬盘驱动器(Hard Disk Drive,HDD)、软盘驱动器、闪存、光盘、磁光盘、磁带或通用串行总线(Universal Serial Bus,USB)驱动器或者两个或更多个以上这些的组合。在合适的情况下,存储器402可包括可移除或不可移除(或固定)的介质。在合适的情况下,存储器402可在数据处理装置的内部或外部。在特定实施例中,存储器402是非易失性固态存储器。在特定实施例中,存储器402包括只读存储器(Read-Only Memory,ROM)。在合适的情况下,该ROM可以是掩模编程的ROM、可编程ROM(Programmable read-only memory,PROM)、可擦除PROM(Erasable Programmable ROM,EPROM)、电可擦除PROM(Electrically Erasable Programmable Read Only Memory EEPROM)、电可改写ROM(EAROM)或闪存或者两个或更多个以上这些的组合。
处理器401通过读取并执行存储器402中存储的计算机程序指令,以实现上述实施例中的任意一种基于C&C域名分析的僵尸网络检测方法。
在一个示例中,计算机设备还可包括通信接口403和总线410。其中,如图4所示,处理器401、存储器402、通信接口403通过总线410连接并完成相互间的通信。
通信接口403,主要用于实现本发明实施例中各模块、装置、单元和/或设备之间的通信。
总线410包括硬件、软件或两者,将计算机设备的部件彼此耦接在一起。举例来说而非限制,总线可包括加速图形端口(Accelerated Graphic Ports或者Advanced Graphic Ports,AGP)或其他图形总线、增强工业标准架构(Extended Industry Standard Architecture,EISA)总线、前端总线(Front Side Bus,FSB)、超传输(HyperTransport,HT)互连、工业标准架构(Industry Standard Architecture,ISA)总线、无限带宽互连、低引脚数(Low Pin Count,LPC)总线、存储器总线、微信道架构(MicroChannel Architecture,MCA)总线、外围组件互连(Peripheral Component Interconnect,PCI)总线、PCI-Express(PCI-X)总线、串行高级技术附件(Serial Advanced Technology Attachment,SATA)总线、视频电子标准协会局部(VESA local bus,VLB)总线或其他合适的总线或者两个或更多个以上这些的组合。在合适的情况下,总线410可包括一个或多个总线。尽管本发明实施例描述和示出了特定的总线,但本发明考虑任何合适的总线或互连。
实施例四
另外,结合上述实施例中的基于C&C域名分析的僵尸网络检测方法,本发明实施例可提供一种计算机可读存储介质来实现。该计算机可读存储介质上存储有计算机程序指令;该计算机程序指令被处理器执行时实现上述实施例中的任意一种基于C&C域名分析的僵尸网络检测方法。
需要明确的是,本发明并不局限于上文所描述并在图中示出的特定配置和处理。为了简明起见,这里省略了对已知方法的详细描述。在上述实施例中,描述和示出了若干具体的步骤作为示例。但是,本发明的方法过程并不限于所描述和示出的具体步骤,本领域的技术人员可以在领会本发明的精神后,作出各种改变、修改和添加,或者改变步骤之间的顺序。
以上所述的结构框图中所示的功能块可以实现为硬件、软件、固件或者它们的组合。当以硬件方式实现时,其可以例如是电子电路、专用集成电路(ASIC)、适当的固件、插件、功能卡等等。当以软件方式实现时,本发明的元素是被用于执行所需任务的程序或者代码段。程序或者代码段可以存储在机器可读介质中,或者通过载波中携带的数据信号在传输介质或者通信链路上传送。“机器可读介质”可以包括能够存储或传输信息的任何介质。机器可读介质的例子包括电子电路、半导体存储器设备、ROM、闪存、可擦除ROM(EROM)、软盘、CD-ROM、光盘、硬盘、光纤介质、射频(RF)链路,等等。代码段可以经由诸如因特网、内联网等的计算机网络被下载。
还需要说明的是,本发明中提及的示例性实施例,基于一系列的步骤或者装置描述一些方法或系统。但是,本发明不局限于上述步骤的顺序,也就是说,可以按照实施例中提及的顺序执行步骤,也可以不同于实施例中的顺序,或者若干步骤同时执行。
以上所述,仅为本发明的具体实施方式,所属领域的技术人员可以清楚地了解到,为了描述的方便和简洁,上述描述的系统、模块和单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。应理解,本发明的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本发明揭露的技术范围内,可轻易想到各种等效的修改或替换,这些修改或替换都应涵盖在本发明的保护范围之内。
尽管本发明已进行了一定程度的描述,明显地,在不脱离本发明的精神和范围的条件下,可进行各个条件的适当变化。可以理解,本发明不限于所述实 施方案,而归于权利要求的范围,其包括所述每个因素的等同替换。

Claims (15)

  1. 一种基于C&C域名分析的僵尸网络检测方法,其特征在于,包括:
    信息获取步骤,获取DNS日志记录;
    域名分析步骤,根据预先构建的域名分析器,检测所述DNS日志记录中的C&C域名,并判断每条C&C域名的所属类别;
    僵尸网络确定步骤,根据所述C&C域名及所述C&C域名的所属类别,确定是否存在僵尸网络。
  2. 根据权利要求1所述的方法,其特征在于,还包括:
    数据统计步骤,统计每类C&C域名的发生频次;
    趋势判断步骤,根据所有类别的C&C域名的发生频次,确定所述僵尸网络的活动趋势,以辅助及时制定有效的抑制措施。
  3. 根据权利要求2所述的方法,其特征在于,所述趋势判断步骤,包括:
    将每类C&C域名的发生频次代入泊松分布概率函数,以获取对应所属类别的泊松参数;
    将所有所述泊松参数确定为僵尸网络活动规律衡量指标;
    根据所述僵尸网络活动规律衡量指标,确定所述僵尸网络的活动趋势。
  4. 根据权利要求1所述的方法,其特征在于,所述域名分析器的训练过程,包括:
    对合法网站公开的合法域名进行清洗以获取合法域名集;
    采用域名生成算法生成C&C域名集,并对所述C&C域名集中的每个域名进行分类标记;
    统计分析所述合法域名集和所述C&C域名集中的每个域名的结构,构建各类域名的字符概率字典;
    从所述合法域名集中随机抽取设定数量的合法域名,获取合法域名训 练样本集;
    从所述C&C域名集中随机抽取设定数量的C&C域名,获取C&C域名训练样本集,
    根据所述合法域名训练样本集、所述C&C域名训练样本集和所述字符概率字典,对所述域名分析器进行训练。
  5. 根据权利要求1或4所述的方法,其特征在于,所述域名分析器为,基于累积BP算法的神经网络模型的计算,且所述神经网络模型中设置有综合考虑经验误差因子和网络复杂度因子的正则化项。
  6. 根据权利要求5所述的方法,其特征在于,所述基于累积BP算法的神经网络模型的计算步骤,包括:
    计算误差目标函数;
    描述神经网络复杂度;
    通过交叉验证法估计模型参数;
    使用随机梯度下降调参逼近误差函数全局最小解。
  7. 根据权利要求1所述的方法,其特征在于,所述域名分析步骤包括:
    提取DNS日志记录中的域名;
    对所述域名进行特征提取;
    根据所述字符概率字典确定所述域名是否为C&C域名;
    对所述C&C域名进行域名特征量化,以获取所述C&C域名的分类号;
    根据所述分类号确定所述C&C域名的所属类别。
  8. 一种基于C&C域名分析的僵尸网络检测装置,其特征在于,包括:
    信息获取单元,用于获取DNS日志记录;
    域名分析单元,用于根据预先构建的域名分析器,检测所述DNS日志记录中的C&C域名,并判断每条C&C域名的所属类别;
    僵尸网络确定单元,用于根据所述C&C域名及所述C&C域名的所属 类别,确定是否存在僵尸网络。
  9. 根据权利要求8所述的基于C&C域名分析的僵尸网络检测装置,其特征在于,还包括:
    数据统计单元,用于统计每类C&C域名的发生频次;
    趋势判断单元,用于根据所有类别的C&C域名的发生频次,确定所述僵尸网络的活动趋势,以辅助及时制定有效的抑制措施。
  10. 根据权利要求9所述的基于C&C域名分析的僵尸网络检测装置,其特征在于,所述趋势判断单元,用于将每类C&C域名的发生频次代入泊松分布概率函数,以获取对应所属类别的泊松参数;将所有所述泊松参数确定为僵尸网络活动规律衡量指标;根据所述僵尸网络活动规律衡量指标,确定所述僵尸网络的活动趋势。
  11. 根据权利要求8所述的基于C&C域名分析的僵尸网络检测装置,其特征在于,所述域名分析器的训练过程,包括:
    对合法网站公开的合法域名进行清洗以获取合法域名集;
    采用域名生成算法生成C&C域名集,并对所述C&C域名集中的每个域名进行分类标记;
    统计分析所述合法域名集和所述C&C域名集中的每个域名的结构,构建各类域名的字符概率字典;
    从所述合法域名集中随机抽取设定数量的合法域名,获取合法域名训练样本集;
    从所述C&C域名集中随机抽取设定数量的C&C域名,获取C&C域名训练样本集,
    根据所述合法域名训练样本集、所述C&C域名训练样本集和所述字符概率字典,对所述域名分析器进行训练。
  12. 根据权利要求8或11所述的基于C&C域名分析的僵尸网络检测 装置,其特征在于,所述域名分析器为,基于累积BP算法的神经网络模型的计算,且所述神经网络模型中设置有综合考虑经验误差因子和网络复杂度因子的正则化项。
  13. 根据权利要求8所述的基于C&C域名分析的僵尸网络检测装置,其特征在于,所述域名分析单元用于:
    提取DNS日志记录中的域名;
    对所述域名进行特征提取;
    根据所述字符概率字典确定所述域名是否为C&C域名;
    对所述C&C域名进行域名特征量化,以获取所述C&C域名的分类号;
    根据所述分类号确定所述C&C域名的所属类别。
  14. 一种计算机设备,其特征在于,包括:至少一个处理器、至少一个存储器以及存储在所述存储器中的计算机程序指令,当所述计算机程序指令被所述处理器执行时实现如权利要求1-7中任一项所述的方法。
  15. 一种计算机可读存储介质,其上存储有计算机程序指令,其特征在于,当所述计算机程序指令被处理器执行时实现如权利要求1-7中任一项所述的方法。
PCT/CN2018/096107 2018-01-15 2018-07-18 基于c&c域名分析的僵尸网络检测方法、装置、设备及介质 WO2019136953A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810036078.7A CN108156174B (zh) 2018-01-15 2018-01-15 基于c&c域名分析的僵尸网络检测方法、装置、设备及介质
CN201810036078.7 2018-01-15

Publications (1)

Publication Number Publication Date
WO2019136953A1 true WO2019136953A1 (zh) 2019-07-18

Family

ID=62461365

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/096107 WO2019136953A1 (zh) 2018-01-15 2018-07-18 基于c&c域名分析的僵尸网络检测方法、装置、设备及介质

Country Status (2)

Country Link
CN (1) CN108156174B (zh)
WO (1) WO2019136953A1 (zh)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112966713A (zh) * 2021-02-02 2021-06-15 杭州安恒信息技术股份有限公司 基于深度学习的dga域名检测方法、装置及计算机设备
CN113158660A (zh) * 2021-04-09 2021-07-23 深圳市联软科技股份有限公司 应用于渗透测试的子域名发现方法及系统
CN113746952A (zh) * 2021-09-14 2021-12-03 京东科技信息技术有限公司 Dga域名检测方法、装置、电子设备及计算机存储介质
CN114257565A (zh) * 2020-09-10 2022-03-29 中国移动通信集团广东有限公司 挖掘潜在威胁域名的方法、系统和服务器
CN114363062A (zh) * 2021-12-31 2022-04-15 深信服科技股份有限公司 一种域名检测方法、系统、设备及计算机可读存储介质
CN114615003A (zh) * 2020-12-07 2022-06-10 中国移动通信有限公司研究院 命令和控制c&c域名的验证方法、装置及电子设备
CN114826758A (zh) * 2022-05-11 2022-07-29 绿盟科技集团股份有限公司 一种针对域名解析系统dns的安全分析方法及装置
CN114866246A (zh) * 2022-04-12 2022-08-05 东莞职业技术学院 基于大数据的计算机网络安全入侵检测方法
CN115134095A (zh) * 2021-03-10 2022-09-30 中国电信股份有限公司 僵尸网络控制端检测方法及装置、存储介质、电子设备
CN115333850A (zh) * 2022-08-26 2022-11-11 中国电信股份有限公司 域名检测方法、系统及相关设备
US11683337B2 (en) * 2020-06-11 2023-06-20 T-Mobile Usa, Inc. Harvesting fully qualified domain names from malicious data packets

Families Citing this family (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11374897B2 (en) 2018-01-15 2022-06-28 Shenzhen Leagsoft Technology Co., Ltd. CandC domain name analysis-based botnet detection method, device, apparatus and medium
CN108156174B (zh) * 2018-01-15 2020-03-27 深圳市联软科技股份有限公司 基于c&c域名分析的僵尸网络检测方法、装置、设备及介质
CN109246074A (zh) * 2018-07-23 2019-01-18 北京奇虎科技有限公司 识别可疑域名的方法、装置、服务器及可读存储介质
CN109246083B (zh) * 2018-08-09 2021-08-03 奇安信科技集团股份有限公司 一种dga域名的检测方法及装置
US10764246B2 (en) * 2018-08-14 2020-09-01 Didi Research America, Llc System and method for detecting generated domain
CN110798439B (zh) * 2018-09-04 2022-04-19 国家计算机网络与信息安全管理中心 主动探测物联网僵尸网络木马的方法、设备及存储介质
CN109977221B (zh) * 2018-09-04 2023-09-19 中国平安人寿保险股份有限公司 基于大数据的用户验证方法及装置、存储介质、电子设备
CN109450845B (zh) * 2018-09-18 2020-08-04 浙江大学 一种基于深度神经网络的算法生成恶意域名检测方法
CN109784049B (zh) * 2018-12-21 2021-04-09 奇安信科技集团股份有限公司 威胁数据处理的方法、设备、系统和介质
CN109617909B (zh) * 2019-01-07 2021-04-27 福州大学 一种基于smote和bi-lstm网络的恶意域名检测方法
CN110149331B (zh) * 2019-05-22 2021-07-06 中国科学院长春光学精密机械与物理研究所 一种P2P botnet检测方法、装置和介质
CN110225030B (zh) * 2019-06-10 2021-09-28 福州大学 基于rcnn-spp网络的恶意域名检测方法及系统
CN112839012B (zh) * 2019-11-22 2023-05-09 中国移动通信有限公司研究院 僵尸程序域名识别方法、装置、设备及存储介质
CN111628970B (zh) * 2020-04-24 2021-10-15 中国科学院计算技术研究所 一种dga型僵尸网络的检测方法、介质和电子设备
CN111953673B (zh) * 2020-08-10 2022-07-05 深圳市联软科技股份有限公司 一种dns隐蔽隧道检测方法及系统
CN112949768A (zh) * 2021-04-07 2021-06-11 苏州瑞立思科技有限公司 一种基于lstm的流量分类方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101741862A (zh) * 2010-01-22 2010-06-16 西安交通大学 基于数据包序列特征的irc僵尸网络检测系统和检测方法
CN103152442A (zh) * 2013-01-31 2013-06-12 中国科学院计算机网络信息中心 一种僵尸网络域名的检测与处理方法及系统
CN106453412A (zh) * 2016-12-01 2017-02-22 绵阳灵先创科技有限公司 一种基于频次特征的恶意域名判定方法
WO2017223342A1 (en) * 2016-06-22 2017-12-28 Ntt Innovation Institute, Inc. Botnet detection system and method
CN108156174A (zh) * 2018-01-15 2018-06-12 深圳市联软科技股份有限公司 基于c&c域名分析的僵尸网络检测方法、装置、设备及介质

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2007050244A2 (en) * 2005-10-27 2007-05-03 Georgia Tech Research Corporation Method and system for detecting and responding to attacking networks
US8533819B2 (en) * 2006-09-29 2013-09-10 At&T Intellectual Property Ii, L.P. Method and apparatus for detecting compromised host computers
CN104580249B (zh) * 2015-01-28 2019-05-07 北京润通丰华科技有限公司 一种基于日志的僵木蠕网络分析方法和系统
CN105072214B (zh) * 2015-08-28 2018-10-09 携程计算机技术(上海)有限公司 基于域名特征的c&c域名识别方法
CN105897714B (zh) * 2016-04-11 2018-11-09 天津大学 基于dns流量特征的僵尸网络检测方法
CN106657001B (zh) * 2016-11-10 2019-12-13 广州赛讯信息技术有限公司 一种基于Netflow及DNS日志的僵尸网络检测方法
CN106549980B (zh) * 2016-12-30 2020-04-07 北京神州绿盟信息安全科技股份有限公司 一种恶意c&c服务器确定方法及装置
CN107196910B (zh) * 2017-04-18 2019-09-10 国网山东省电力公司电力科学研究院 基于大数据分析的威胁预警监测系统、方法及部署架构
CN107404473A (zh) * 2017-06-06 2017-11-28 西安电子科技大学 基于Mshield机器学习多模式Web应用防护方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101741862A (zh) * 2010-01-22 2010-06-16 西安交通大学 基于数据包序列特征的irc僵尸网络检测系统和检测方法
CN103152442A (zh) * 2013-01-31 2013-06-12 中国科学院计算机网络信息中心 一种僵尸网络域名的检测与处理方法及系统
WO2017223342A1 (en) * 2016-06-22 2017-12-28 Ntt Innovation Institute, Inc. Botnet detection system and method
CN106453412A (zh) * 2016-12-01 2017-02-22 绵阳灵先创科技有限公司 一种基于频次特征的恶意域名判定方法
CN108156174A (zh) * 2018-01-15 2018-06-12 深圳市联软科技股份有限公司 基于c&c域名分析的僵尸网络检测方法、装置、设备及介质

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11683337B2 (en) * 2020-06-11 2023-06-20 T-Mobile Usa, Inc. Harvesting fully qualified domain names from malicious data packets
CN114257565B (zh) * 2020-09-10 2023-09-05 中国移动通信集团广东有限公司 挖掘潜在威胁域名的方法、系统和服务器
CN114257565A (zh) * 2020-09-10 2022-03-29 中国移动通信集团广东有限公司 挖掘潜在威胁域名的方法、系统和服务器
CN114615003A (zh) * 2020-12-07 2022-06-10 中国移动通信有限公司研究院 命令和控制c&c域名的验证方法、装置及电子设备
CN112966713A (zh) * 2021-02-02 2021-06-15 杭州安恒信息技术股份有限公司 基于深度学习的dga域名检测方法、装置及计算机设备
CN112966713B (zh) * 2021-02-02 2024-03-19 杭州安恒信息技术股份有限公司 基于深度学习的dga域名检测方法、装置及计算机设备
CN115134095A (zh) * 2021-03-10 2022-09-30 中国电信股份有限公司 僵尸网络控制端检测方法及装置、存储介质、电子设备
CN113158660B (zh) * 2021-04-09 2023-03-21 深圳市联软科技股份有限公司 应用于渗透测试的子域名发现方法及系统
CN113158660A (zh) * 2021-04-09 2021-07-23 深圳市联软科技股份有限公司 应用于渗透测试的子域名发现方法及系统
CN113746952A (zh) * 2021-09-14 2021-12-03 京东科技信息技术有限公司 Dga域名检测方法、装置、电子设备及计算机存储介质
CN113746952B (zh) * 2021-09-14 2024-04-16 京东科技信息技术有限公司 Dga域名检测方法、装置、电子设备及计算机存储介质
CN114363062A (zh) * 2021-12-31 2022-04-15 深信服科技股份有限公司 一种域名检测方法、系统、设备及计算机可读存储介质
CN114866246A (zh) * 2022-04-12 2022-08-05 东莞职业技术学院 基于大数据的计算机网络安全入侵检测方法
CN114866246B (zh) * 2022-04-12 2023-07-04 东莞职业技术学院 基于大数据的计算机网络安全入侵检测方法
CN114826758A (zh) * 2022-05-11 2022-07-29 绿盟科技集团股份有限公司 一种针对域名解析系统dns的安全分析方法及装置
CN114826758B (zh) * 2022-05-11 2023-05-16 绿盟科技集团股份有限公司 一种针对域名解析系统dns的安全分析方法及装置
CN115333850A (zh) * 2022-08-26 2022-11-11 中国电信股份有限公司 域名检测方法、系统及相关设备
CN115333850B (zh) * 2022-08-26 2024-04-23 中国电信股份有限公司 域名检测方法、系统及相关设备

Also Published As

Publication number Publication date
CN108156174A (zh) 2018-06-12
CN108156174B (zh) 2020-03-27

Similar Documents

Publication Publication Date Title
WO2019136953A1 (zh) 基于c&c域名分析的僵尸网络检测方法、装置、设备及介质
US11374897B2 (en) CandC domain name analysis-based botnet detection method, device, apparatus and medium
CN109951500B (zh) 网络攻击检测方法及装置
US10721244B2 (en) Traffic feature information extraction method, traffic feature information extraction device, and traffic feature information extraction program
US9794229B2 (en) Behavior analysis based DNS tunneling detection and classification framework for network security
CN106657001B (zh) 一种基于Netflow及DNS日志的僵尸网络检测方法
US8549645B2 (en) System and method for detection of denial of service attacks
US10721245B2 (en) Method and device for automatically verifying security event
CN107770132B (zh) 一种对算法生成域名进行检测的方法及装置
US10944784B2 (en) Identifying a potential DDOS attack using statistical analysis
JP2019021294A (ja) DDoS攻撃判定システムおよび方法
WO2015047803A1 (en) Malware detection based on vm behavioral analysis and machine learning classification
Hu et al. BAYWATCH: robust beaconing detection to identify infected hosts in large-scale enterprise networks
JP5832951B2 (ja) 攻撃判定装置、攻撃判定方法及び攻撃判定プログラム
CN107209834B (zh) 恶意通信模式提取装置及其系统和方法、记录介质
CN112437062B (zh) 一种icmp隧道的检测方法、装置、存储介质和电子设备
CN106911665B (zh) 一种识别恶意代码弱口令入侵行为的方法及系统
US10721148B2 (en) System and method for botnet identification
CN114785567A (zh) 一种流量识别方法、装置、设备及介质
CN111885034B (zh) 物联网攻击事件追踪方法、装置和计算机设备
Niu et al. Using XGBoost to discover infected hosts based on HTTP traffic
Chiba et al. Botprofiler: Profiling variability of substrings in http requests to detect malware-infected hosts
CN113938312B (zh) 一种暴力破解流量的检测方法及装置
CN117391214A (zh) 模型训练方法、装置及相关设备
Kheir et al. Peerviewer: Behavioral tracking and classification of P2P malware

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18900319

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 12.03.2021.)

122 Ep: pct application non-entry in european phase

Ref document number: 18900319

Country of ref document: EP

Kind code of ref document: A1