Malware detection system with contextual analysis

Description

FIELD

Embodiments of the disclosure relate to the field of cyber-security. More specifically, one embodiment of the disclosure relates to a system, apparatus and method configured to detect malicious objects based on an analysis of contextual information.

GENERAL BACKGROUND

Over the last decade, malicious software has become a pervasive problem for Internet users as many networked resources include vulnerabilities that are subject to attack. For instance, over the past few years, an increasing number of vulnerabilities are being discovered in software that is loaded onto network devices, such as vulnerabilities within operating systems, for example. While some vulnerabilities continue to be addressed through software patches, prior to the release of such software patches, network devices will continue to be targeted for attack by malware, namely information such as computer code that attempts during execution to take advantage of a vulnerability in computer software by acquiring sensitive information or adversely influencing or attacking normal operations of the network device or the entire enterprise network.

Moreover, with the proliferation of the Internet and the reliance on electronic email (email) as a means of communication, malware is capable of spreading more quickly and affecting a larger subset of the population than ever before.

Although some conventional malware detection systems may be configured to evaluate objects for malware, these conventional systems may produce “false negative” or “false positive” outcomes because the classification of the objects is based on a collection of scores that concentrate on results associated with a single type of analysis. For example, according to conventional malware detection systems, the results from a static analysis of an object (analysis of characteristics of the object) are correlated and subsequently assigned a first score while results from a behavioral analysis of the object (analysis of behaviors associated with the object during processing) are correlated and subsequently assigned a second score. The first and second scores are evaluated to classify the object as malicious or non-malicious.

As a result, in accordance with conventional malware detection systems, an object may contain features that, when analyzed in isolation, may fail to identify the object as malicious when the feature itself does not exhibit maliciousness during the scanning process. This “false negative” tends to occur for objects that are part of a greater, multi-stage attack. Consequently, the malicious object may be allowed to pass through to the end user. It is contemplated that conventional malware detection systems may also experience a “false positive” outcome since there is no correlation of results from different analyses of the object to determine whether the object is malicious.

Accordingly, a need exists for an improved malware detection system, apparatus and method that further mitigates inaccurate outcomes concerning malware detection.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention are illustrated by way of example and not by way of limitation in the figures of the accompanying drawings, in which like references indicate similar elements and in which:

FIG. 1A is an exemplary block diagram of an operational flow of an malware detection system within a network device.

FIG. 1B is a logical representation of the classification engine in communication with the static analysis logic system, the dynamic (behavioral) analysis logic system, and optionally the emulation logic system of FIG. 1A.

FIG. 1C is another logical representation of the classification engine of FIG. 1B in feedback communications with a controller for initiating secondary contextual analyses.

FIG. 2 is an exemplary block diagram of a communication system deploying the malware detection system of FIG. 1 via a network.

FIG. 3 is an exemplary embodiment of a logical representation of the malware detection system of FIG. 2.

FIG. 4 is an exemplary flowchart of a contextual analysis of attributes associated with an object obtained through a plurality of analyses.

FIG. 5 is a flowchart of an exemplary method for applying threat index values on an observed multi-type attribute combination with characteristics, behaviors and/or emulated results during correlation and classification of the object.

FIG. 6 is an exemplary network device demonstrating an alert according to the present disclosure.

DETAILED DESCRIPTION

Embodiments of the present disclosure generally relate to a malware detection system configured to detect and classify one or more objects based on a correlation of analytic results that are produced from different types of analyses conducted on that object. According to one embodiment of the disclosure, these analytic results, generally referred to herein as “attributes,” may feature (1) one or more (a first set of) behaviors that include monitored activity or inactivity by the object during processing (herein, “behavior(s)”) and (2) one or more (a second set of) characteristics of the object that include certain information pertaining to the object acquired without executing or opening the object (herein, “characteristic(s)”). Examples of types of monitored behaviors may include communication-based or execution-based activities (especially anomalies) while types of detected characteristics may include metadata, any formatting information (especially anomalies) of the object, and/or a score associated with the behavior(s) or characteristic(s). Of course, it is contemplated that these analytic results may further include results obtained from emulated operations on the object (herein, “emulated result(s)”).

Specifically, the malware detection system features a plurality of analysis logic systems. Communicatively coupled to attribute correlation logic of a classification engine, each of the analysis logic systems separately analyzes different aspects of an object under analysis. Such analyses produce attributes and these attributes are provided to the attribute correlation logic. Operating in accordance with an extended correlation rule set, the attribute correlation logic may analyze the received attributes not only for the presence or absence of certain attributes, but also the presence or absence of certain combinations of attributes, including one or more combination of attributes received from different analysis logic systems (herein referred to as “multi-type attribute combinations”). The different analysis logic systems may include at least an analysis logic system that involves processing (execution) of the object (e.g., dynamic analysis) and an analysis logic system that scans for content associated with the object without its opening or execution (e.g., static analysis). The attributes and certain combinations of attributes constitute contextual information that is used for determining threat indices, which are relied upon in determining whether or not the object is malicious.

Responsive to receiving contextual information from the attribute correlation logic, threat index generation logic determines a threat index associated with the object. The threat index may be used to represent, at least in part, (i) a score that corresponds to the likelihood of the object being malicious (e.g., object includes malware or is part of a malicious attack); (ii) a policy violation involving a detected presence of a certain type of object (e.g., receipt of a particular type of document including macros, etc.); (iii) a severity of a potential malicious attack associated with the object; or (iv) an ordering (priority) for secondary contextual analysis (as needed).

According to one embodiment of the disclosure, each attribute (e.g., each characteristic, behavior, and/or emulated result) provided from the plurality of analysis logic systems may be associated with a threat index value. Additionally, or in the alternative, multi-type attribute combinations (i.e., combinations of attributes provided from the different analysis logic systems) may be assigned threat index values. The threat index for the object may be produced as an aggregate of these threat index values or as a function of some or all of these index values. As a result, object classification is accomplished by determining a threat index (level) of the object that is based, at least in part, on the existence of one or more selected groups of attributes from different types of analyses (e.g., one or more characteristics and one or more behaviors) instead of simply relying on discrete correlations of analytic results from the static analysis and the analytic results from the dynamic analysis. This consideration of multi-type attribute combinations for object classification increases malware detection accuracy, which reduces the number of “false negative” results as well as “false positive” results.

Additionally, or in the alternative, the above-described malware detection involves a correlation of attributes across one or more vectors. A “vector” includes information that is directed to the delivery mechanism of the object. For instance, as an illustrative example, for a vector based on network (packet-based) communications, the attributes may include information extracted from a header of a packet such as a source address (e.g., Internet Protocol “IP” address, domain name, etc.) or a destination address. As another illustrative example, for a vector based on email communications, the attributes may include information within the body or header of the email message, such as information associated with the “From” field (e.g., the email sender's information/address), the “To” field (e.g., the email recipient's information/address; the “Subject” line, or the like.

The correlation of the attributes across one or more vectors may improve accuracy in the classification of the object. Also, it is further contemplated that cross vector analysis may be particularly useful in detecting multi-stage attacks, which may have components delivered to a targeted network device across multiple vectors in an attempt to infiltrate the targeted network device.

Additionally, or in the alternative, the classification engine may further prompt a secondary contextual analysis after a contextual analysis of the attributes (e.g., monitored behavior(s) and/or detected characteristic(s)) has been conducted. The secondary contextual analysis may correspond to a “hunting” for particular attributes, such as particular characteristic(s) of the object (as determined by a static analysis logic system) and/or particular behavior(s) of the object (as determined by a dynamic analysis logic system) for example, which may increase or decrease the threat index currently associated with the object. Hence, using the secondary contextual analysis, a more robust object classification scheme is provided for determining whether the object under analysis is malicious or non-malicious.

Herein, according to one embodiment of the disclosure, the secondary contextual analysis may be triggered by an event whose occurrence may want a more in-depth analysis to mitigate potential “false negative” and “false positive” outcomes. The event may simply include a detection of the object under analysis being suspicious (i.e., neither definitively malicious nor definitively non-malicious (benign)). For instance, in response to the one or more monitored behaviors being determined to provide a “borderline” threat index that the object under analysis is malicious (e.g., barely exceeds a malicious threshold), the classification engine may examine particular attributes (e.g., certain characteristics provided by the static analysis logic system). The particular characteristic(s) may be selected based on one or more correlation rules provided to the classification engine, where these rules now include targeted analysis associated with combinations of behavior(s) and/or characteristic(s) associated with previous determined malicious objects. Also, the presence of the particular characteristic(s) provide further evidence that the object is malicious (or non-malicious), which may cause the threat index associated with the object to either increase to more definitely identify that the object is malicious (reduce “false negative” outcomes) or decrease to now fall under the threshold (reduce “false positive” outcomes).

Additionally, in order to avoid a false positive verdict, the secondary contextual analysis may be triggered in response to an observation by an analysis logic system of the object conducting a malicious activity that may not have been fully considered in light of other characteristics or behaviors. Also, it is further contemplated that the secondary contextual analysis may be triggered based on a particular behavior conducted by the object during processing which is commonly, based on experiential knowledge or machine learning, associated with particular observable characteristics that were not identified during the prior contextual analysis.

Once maliciousness has been confirmed with regard to the object, an alert is generated. The alert (e.g., a type of messages including text message or email message, a transmitted displayable image, or other types of information transmitted over a wired or wireless communication path) warns a network administrator that an analyzed object has been detected by the malware detection system as malicious. The alert may include the threat index for the object under analysis and some or all of its observed attributes, and in some embodiments, a signature and contextual information that identifies the detected malware attack type, including without limitation, phishing campaigns, Advanced Persistent Threats (APT), Point-Of-Sales attacks (POS), Crimeware attacks, or the like.

Terminology

In the following description, certain terminology is used to describe features of the invention.

In certain situations, both terms “logic,” “logic system,” “engine” and “component” are representative of hardware, firmware and/or software that is configured to perform one or more functions. As hardware, logic (or engine or component) may include circuitry having data processing or storage functionality. Examples of such circuitry may include, but are not limited or restricted to a microprocessor, one or more processor cores, a programmable gate array, a microcontroller, an application specific integrated circuit, wireless receiver, transmitter and/or transceiver circuitry, semiconductor memory, or combinatorial logic.

Logic (or logic system/engine/component) may be software in the form of one or more software modules, such as executable code in the form of an executable application, an application programming interface (API), a subroutine, a function, a procedure, an applet, a servlet, a routine, source code, object code, a shared library/dynamic load library, or one or more instructions. These software modules may be stored in any type of a suitable non-transitory storage medium, or transitory storage medium (e.g., electrical, optical, acoustical or other form of propagated signals such as carrier waves, infrared signals, or digital signals). Examples of non-transitory storage medium may include, but are not limited or restricted to a programmable circuit; a semiconductor memory; non-persistent storage such as volatile memory (e.g., any type of random access memory “RAM”); persistent storage such as non-volatile memory (e.g., read-only memory “ROM”, power-backed RAM, flash memory, phase-change memory, etc.), a solid-state drive, hard disk drive, an optical disc drive, or a portable memory device. As firmware, the executable code is stored in persistent storage.

Hence, the term “computerized” generally represents that any corresponding operations are conducted by hardware in combination with software and/or firmware.

The term “object” generally refers to a collection of data, whether in transit (e.g., over a network) or at rest (e.g., stored), often having a logical structure or organization that enables it to be classified for purposes of analysis. During analysis, for example, the object may exhibit a set of expected characteristics and, during processing, a set of expected behaviors. The object may also exhibit a set of unexpected characteristics and a set of unexpected behaviors that may evidence a potential presence of malware and allow the object to be classified as malicious.

Examples of objects may include one or more flows or one or more self-contained elements. A “flow” generally refers to related packets that are received, transmitted, or exchanged within a communication session. For convenience, a packet is broadly referred to as a series of bits or bytes having a prescribed format, which may include packets, frames, or cells. One example of a flow is an electronic mail (email) message that includes related packets received as a transmission.

As another illustrative example, an object may include a set of flows such as (1) a sequence of transmissions in accordance with a particular communication protocol (e.g., User Datagram Protocol (UDP); Transmission Control Protocol (TCP); or Hypertext Transfer Protocol (HTTP); or Simple Mail Transfer Protocol (SMTP); or Internet Message Access Protocol (IMAP); or Post Office Protocol (POP)), or (2) inter-process communications (e.g., Remote Procedure Call “RPC” or analogous processes, etc.). Moreover, the object may be featured with a message as an attachment, for example.

Similarly, as another illustrative example, the object may be one or more self-contained elements, which may include an attachment to a message. The object, as one or more self-contained elements, may include, but is not limited or restricted to, an executable file, non-executable file (such as a document or a dynamically link library), a Portable Document Format (PDF) file, a JavaScript file, Zip file, a Flash file, a document (for example, a Microsoft Office® document), an electronic mail (email), a HyperText Markup Language (HTML) file representing a downloaded web page, an instant messaging element in accordance with Session Initiation Protocol (SIP) or another messaging protocol, or the like. The object may also refer to any of various files, including, but not limited to executable programs (file types: .com, .exe, .vbs, .zip, .scr, .dll, .pif, .js), macros (file types: .doc, .dot, .pdf, .ppt, .xls, .xlt), embedded scripts, Uniform Resource Locators (URLs), and the like. The object is made available to the malware detection system as data itself or a reference (e.g., pointer, file path, etc.) to a storage location of the object.

According to one embodiment, the term “malware” may be construed broadly as any code or activity that initiates a malicious attack and/or operations associated with anomalous or unwanted behavior. For instance, malware may correspond to a type of malicious computer code that executes an exploit to take advantage of a vulnerability, for example, to harm or co-opt operation of a network device or misappropriate, modify or delete data. In the alternative, malware may correspond to an exploit, namely information (e.g., executable code, data, command(s), etc.) that attempts to take advantage of a vulnerability in software and/or an action by a person gaining unauthorized access to one or more areas of a network device to cause the network device to experience undesirable or anomalous behaviors. The undesirable or anomalous behaviors may include a communication-based anomaly or an execution-based anomaly, which, for example, could (1) alter the functionality of a network device executing application software in an atypical manner (a file is opened by a first process where the file is configured to be opened by a second process and not the first process); (2) alter the functionality of the network device executing that application software without any malicious intent; and/or (3) provide unwanted functionality which may be generally acceptable in another situation. Additionally, malware may be code that initiates unwanted behavior which may be, as one example, uploading a contact list from an endpoint device to cloud storage without receiving permission from the user.

In certain instances, the term “detected” is used herein to represent that there is a prescribed level of confidence (or probability) of an occurrence. The terms “compare” or “comparison” generally mean determining if a match (e.g., a certain level of correlation) is achieved between two items where one of the items may include a particular data pattern.

The term “network device” should be construed as any electronic device with the capability of connecting to a network. Such a network may be a public network such as the Internet or a private network such as a wireless data telecommunication network, wide area network, a type of local area network (LAN), or a combination of networks. Examples of a network device may include, but are not limited or restricted to, a laptop, a mobile phone, a tablet, a computer, etc.

Finally, the terms “or” and “and/or” as used herein are to be interpreted as inclusive or meaning any one or any combination. Hence, “X, Y or Z” or “X, Y and/or Z” mean “any of the following: X; Y; Z; X and Y; X and Z; Y and Z; X, Y and Z.” An exception to this definition may occur when a combination of elements, functions, steps or acts are in some way inherently mutually exclusive.

As this invention is susceptible to embodiments of many different forms, it is intended that the present disclosure is to be considered as an example of the principles of the invention and not intended to limit the invention to the specific embodiments shown and described.

II. Object Analysis Methodology

A. General Architecture of the Malware Detection System

FIG. 1A is an exemplary block diagram of the general operational flow of a malware detection system 100. The malware detection system 100 includes an extraction logic 110, a plurality of analysis logic systems 120₁-120_N (N≥1; N=3 in FIG. 1A), a classification engine 130 and reporting engine 150. As shown, the extraction logic 110 receives an object 160 via a network 165, although it is contemplated that the object 160 may be uploaded to the malware detection system 100 directly from a peripheral source (e.g., off-line storage device, portable storage device, etc.). The extraction logic 110 is configured to gather meta information associated with the object 160. The meta information may include meta data 161 that may be used to configure one or more virtual machines (not shown) within a dynamic analysis logic system 120₂. Additionally, or in an alternative, the meta information may further include meta data 162 provided to the classification engine 130 for use, at least in part, in classifying the object 160 as malicious or non-malicious.

Thereafter, some or all of the analysis logic systems 120₁-120_N, such as the static analysis logic system 120₁ and the dynamic analysis logic system 120₂ for example, may be configured to receive information associated with the object 160 for analysis. During or after analysis, the analysis logic systems 120₁-120_N provide analytic results (also referred to as “attributes”) to the classification engine 130. The attributes 170₁-170_N from the respective analysis logic systems 120₁-120_N may be correlated and classified by the classification engine 130 for determining whether the object 160 is malicious.

More specifically, based on an expanded configuration rule set 135, the classification engine 130 may be configured for determining whether or not the object 160 should be classified as malicious. As shown in FIG. 1B, the expanded configuration rule set 135 includes an expanded correlation rule set 136 and an expanded threat index set 137. The expanded correlation rule set 136 identifies attribute patterns relied upon for malware detection, where these attribute patterns now include patterns of attribute combinations from different types of analyses (e.g., an attribute pattern including one or more behaviors and one or more characteristics; an attribute pattern including one or more behaviors and one or more emulated results; etc.). The presence or absence of these combinations of attributes from different types of analyses (herein, “multi-type attribute combination”) is used in determining whether the object 160 is malicious or non-malicious. The expanded threat index set 137 includes threat indices associated with the attribute patterns, including one or more attribute patterns associated with a combination of attributes from different types of analyses (herein, “multi-type attribute pattern”).

Referring still to FIG. 1B, a logical representation of the classification engine 130 based on received attributes from the static analysis logic system 120₁, the dynamic (behavioral) analysis logic system 120₂, and optionally the emulation logic system 120₃ is shown. Herein, the object 160 may be analyzed by static analysis logic system 120₁, which determines characteristics of the object 160 such as formatting or patterns of content for the object 160, and provide such characteristics 170₁ to the classification engine 130. The attributes (characteristics) 170₁ provided from the static analysis logic system 120₁ to the classification engine 130 are illustrated as four characteristics referenced as C1, C2, C3 & C4.

Additionally, the dynamic analysis logic system 120₂ may receive object 160, perhaps concurrently with the static analysis logic system 120₁ or from the static analysis logic system 120₂ after analysis. Herein, the dynamic analysis logic system 120₂ may monitor for one or more anomalous (or undesirable) behaviors by a virtual machine (not shown) during its processing of the object 160. The anomalous (or undesirable) behaviors 170₂ may include communication-based anomalies and/or execution-based anomalies, described below. These attributes (behaviors) 170₂ provided from the dynamic analysis logic system 120₂ to the classification engine 130 are illustrated as three monitored behaviors referenced as B1, B2 & B3.

Lastly, as an optional logic system, the emulation logic system 120₃ may receive the object 160, concurrently with or subsequent to another analysis. Herein, the emulation logic system 120₃ may emulate functionality of a targeted client device that conducts operations on the object 160 to produce one or more emulated results 170₃. As shown, in this illustrative embodiment, no attributes 170₃ are provided from the emulation logic system 120₃ to the classification engine 130. However, it is contemplated that, when in operation, emulation logic system 120₃ may provide one or more emulated results (E1 . . . ) to the classification engine 130.

As further shown in FIG. 1B, the classification engine 130 includes attribute correlation logic 140, threat index generation logic 142 and object classification logic 144. Herein, the attribute correlation logic 140 is configured to receive attributes 170₁, 170₂ and/or 170₃ from logic systems 120₁, 120₂ and/or 120₃ respectively. The attribute correlation logic 140 attempts to correlate some or all of the attributes 170₁-170₃ associated with the object 160 in accordance with the expanded correlation rule set 136, which is stored in correlation rules data store 146 (and can be updated as part of configuration rules 135). For this embodiment, the correlation determines what particular attributes and/or combination of attributes, including multi-type attribute combinations, have been collectively detected by the static analysis logic system 120₁ and dynamic analysis logic system 120₂ in accordance with the attribute patterns set forth in the expanded correlation rule set 136.

Herein, as a non-limiting illustration, the attributes and/or combinations of attributes constitute contextual information 141 associated with the object 160, which are provided to the threat index generation logic 142 to determine one or more threat indices 143. The operability of the threat index generation logic 142 is controlled by the expanded threat index set 137, which are stored in index data store 148 and can be updated as part of the configuration rules 135. The one or more threat indices 143 are used by the objection classification logic 144 to determine whether or not the object 160 is malicious.

As shown for illustrative purposes in FIG. 1B, the contextual information 141 includes attributes [C1] and [B2]. The contextual information 141 further includes combinations of attributes from the same analysis type ([C1, C2] or [B1, B3]) as well as combination of attributes from different analysis types, referred to as “multi-type attribute combinations” ([C3, B1, B3] or [C2, B1, B2]). The contextual information 141 corresponds to correlated analytic results based on some or all of the attribute patterns set forth in the expanded correlation rule set 136.

Based on the contextual information 141, the threat index generation logic 142 produces threat index values 143 for these attributes and combinations of attributes: W[C1], W[B2], W[C2, C3], W[B1, B3], W[C3, B1, B3] and/or W[C2, B1, B2]. Based on these threat index values 143, the object classification logic 144 determines whether the object 160 is malicious or not, and provides results 175 associated with the analysis (e.g., some or all of the context information 141, meta information, threat index for object, etc.) to the reporting engine 150.

Referring to FIG. 1C, another illustrative example of the functionality the logic systems 120₁-120_N and the classification engine 130 is shown. Herein, the object 160 may be analyzed by one or more analysis logic systems, including at least a first analysis logic system 120₁ (e.g., static analysis logic system 120₁) and a second analysis logic system 120₂ (e.g., dynamic analysis logic system 120₂). As similarly illustrated in FIG. 1B, the first analysis logic system 120₁ may provide attributes 170₁ (referred as C1, C2, C3 & C4) to the attribute correlation logic 140 of the classification engine 130 while the second analysis logic system 120₂ may provide attributes 170₂ (referred as B1, B2 & B3) to the attribute correlation logic 140.

Herein, the provisioning of one or more virtual machines within the second analysis logic system 120₂ may be conducted by a controller 190. According to one embodiment, the provisioning of the one or more virtual machines within the second analysis logic system 120₂ may be initiated at least in part on meta data 149 associated with the object 160 and, in certain situations, feedback signaling 195 from the object classification logic 144. The controller 190 may provide location information for one or more software profiles (e.g., pointer, memory address, path, etc.) via communication path 1982, which enables re-provisioning of the one or more virtual machines within the second analysis logic system 120₂.

According to one embodiment of the disclosure, the feedback signaling 195 may include information identifying an event that may cause the controller 190 to re-provision one or more virtual machines within the second analysis logic system 120₂ with another software profile (e.g., another version, new software profile directed to processing a different object type, etc.). For instance, as an illustrative example, during processing of the object 160 within one or more virtual machines (not shown) of the second analysis logic system 120₂, the object 160 may create another object (hereinafter “child object”), which constitutes one or more anomalous behaviors (e.g., behavior B1). During classification, the object classification logic 144 detects the presence of the child object, and in response, may return location information (e.g., memory address, pointer, file path, etc.) for the child object to the controller 190 via the feedback signaling 195. The feedback signaling 195 provides the controller 190 with information to conduct one or more additional analyses on the child object (e.g., additional static analysis to obtain characteristics that may assist in the provisioning of the one or more virtual machines within the second analysis logic system 120₂ to subsequently process the child object).

Additionally, the attribute correlation logic 140 attempts to correlate some or all of the attributes 170₁-170₂ associated with the object 160 in accordance with attribute patterns of the expanded correlation rule set 136. For this embodiment, the attribute correlation logic 140 determines whether particular attributes and combination of attributes, including multi-type attribute combinations, have been collectively detected by the static analysis logic system 120₁ and dynamic analysis logic system 120₂. However, according to this illustrative embodiment, one or more attributes (e.g., behavior B3) has not been detected so that multi-type attribute pattern [C3, B1, B3] has not been fully detected.

In certain situations (e.g., low workload, importance of the attribute pattern to the overall threat index computation for the object 160, etc.), the attribute correlation logic 140 may return feedback signaling 197 to the controller 190. In response to receipt of the feedback signaling 197, as represented by communication path 198₂, the controller 190 may coordinate the scheduling (and even provisioning) of virtual machines (not shown) within the second analysis logic system 120₂ to re-process the object 160 to focus monitoring operations on a presence of the attribute B3.

Stated differently, the attribute correlation logic 140 is configured to issue the feedback signaling 197 to controller 190, which commences one or more secondary contextual analyses via communication paths 198₁-198_N to confirm the presence or absence of one or more particular attributes of the object 160 have not been detected by the analysis logic systems 120₁ and 120₂. The secondary contextual analyses may be conducted to mitigate both false negative and false positive outcomes by improving accuracy of the malware detection analysis.

For instance, as an illustrative example, in response to receiving attributes 170₁-170₂ during static and dynamic analysis of the object 160, the attribute correlation logic 140 and threat index generation logic 142 may collectively monitor whether the object 160 has a strong likelihood of being categorized as benign (e.g., collective threat index “0” or “1”), malicious (e.g., collective threat index “9” or “10”) or suspicious (e.g. a collective mid-range threat index score ranging from “2” to “8”). When the object 160 is determined to be suspicious, the attribute correlation logic 140 may provide the feedback signaling 197 in response to one or more particular attributes, which may be set forth in the one or more multi-type attribute patterns of the expanded correlation rule set 136, being undetected. Thereafter, a confirmed presence or absence of any of these particular attributes may enable the object classification logic 144 to conclude, with greater certainty, that the object 160 should be categorized as malicious or non-malicious.

According to another embodiment of the disclosure, as described above, where the object 160 is determined to be highly malicious, the object classification logic 144 may return the feedback signaling 195 to signal the controller 190 to initiate a secondary contextual analysis to confirm that one or more particular attributes or combinations of attributes, which tends to have a larger influence on the classification of the object 160 than other attributes, have been detected. The secondary contextual analysis may be conducted to avoid a false positive result.

According to yet another embodiment of the disclosure, where the object 160 is determined to be associated with a particular vector based on received meta data 162 and/or with a particular characteristic as received from static analysis logic system 120₁, the classification engine 130 may conduct the secondary contextual analysis to assign higher weighting to one or more particular activities conducted by the object 160 during analysis by the dynamic analysis logic system 120₂.

Referring back to FIG. 1A, the classification engine 130 provides the results 175 of its analysis to reporting engine 150 for the purposes of reporting, if appropriate, and generating an alert 180. Herein, the alert 180 is a message that is responsible for warning a targeted entity (e.g., a network administrator) that the object 160 has been detected by the malicious detection system 100 as malicious. The alert 180 may include contextual information that identifies the object and detected malware attack type (e.g., object type, object name, source, destination, or timestamp applied when received over a network) as well as a threat index associated with the object. The threat index may be portrayed, for example as a number provided with other contextual information or an alteration of the other contextual information such as different font color is assigned for different threat index levels (e.g., dark red, 9-10 light red 8-9; orange 7-8, etc.), different font sizes, different image sizes, different positioning of the contextual information on the screen (at top, different area of the screen, different screen, etc.), or other different visual distinguishable features.

Referring now to FIG. 2, an exemplary block diagram of a communication system 200 that features the malware detection system 100 of FIGS. 1A-1C communicatively coupled to the network 165 via an optional firewall 205 and an optional network interface 210 (represented by dashed lines) is shown. Communications between a server device 215 and a client device 220 may be analyzed by the malware detection system 100.

According to the embodiment illustrated in FIG. 2, the malware detection system 100 may be configured as a network device that is adapted to analyze information associated with network traffic routed over the network 165 between at least one server device 215 and at least one client device 220. The network 165 may include a public network such as the Internet, in which case the optional firewall 205 may be interposed on the communication path between the public network and the client device 220. Alternatively, the network 165 may be a private network such as a wireless data telecommunication network, a wide area network, any type of local area network (e.g., LAN, WLAN, etc.), or a combination of networks.

As shown, the malware detection system 100 may be communicatively coupled with the network 165 via a network interface 210. In general, the network interface 120 operates as a data capturing device (sometimes referred to as a “tap” or “network tap”) that is configured to receive data propagating to/from the client device 220 and provide at least some of this data to the malware detection system 100. Alternatively, the malware detection system 100 may be positioned as part of or behind the firewall 115 and in-line with client device 220.

According to one embodiment of the disclosure, the network interface 210 is capable of receiving and routing objects associated with network traffic to the malware detection system 100. The network interface 210 may provide the entire traffic or a certain subset of the network traffic, for example, such as one or more files that are part of a set of flows, packet payloads, or the like. In some embodiments, although not shown, network interface 210 may be contained within the malware detection system 100.

As shown in FIG. 2, the malware detection system 100 includes the extraction logic 110, the static analysis logic system 120₁, the dynamic analysis logic system 120₂, the emulation logic system 120₃, the classification engine 130, and/or the reporting engine 150, a similar architecture as shown in FIG. 1A. Although some of the components disposed within the malware detection system 100 are shown in a communicatively coupled serial configuration, it is envisioned that other topologies may also be implemented, such as, for example, parallel and daisy-chain configurations. It should be appreciated that the extraction logic 110, the static analysis logic system 120₁, the dynamic analysis logic system 120₂, the emulation logic system 120₃, the classification engine 130, and the reporting engine 150 may each be separate and distinct components, but the combination of components may also be implemented in a single block and/or core.

Once the object 160 is captured from the network traffic, it is communicated to the extraction logic 110 of the malware detection system 100. Herein, the extraction logic 110 comprises parsing logic 230, which is configured to extract meta information associated with the object 160. According to one embodiment, the meta information includes meta data 161 that may be used, at least in part by a virtual machine monitor 250, for provisioning one or more virtual machines 255 in the dynamic analysis logic system 120₂. The one or more virtual machines (VMs) 255 conduct run-time processing of at least some of the information associated with the object 160. The meta data 161 may include data directed to the object type (e.g., PDF file, word processing document, HTML (web page) file, etc.), the type of operating system at the source that provided the object 160, web browser type, or the like.

Additionally, or in an alternative, the meta information may further include meta data 162 that may be provided to the classification engine 130. The meta data 162 may be used, at least in part, for classifying the object 160. For example, although not shown in detail, the meta data 162 may be directed to the delivery mechanism (message) associated with the object 160 which, depending on the object type, may include information extracted from a header of a packet (e.g., source IP address, destination IP address, etc.) or from the body or header of the email message (e.g., sender's email address, recipient's email address, subject line, etc.). Hence, although not shown in detail, the meta data 162 may be construed as another analysis type similar to the static analysis (characteristics), dynamic analysis (behaviors), and emulation (e.g., emulation results).

Similarly, it should be appreciated that aspects of the object 160 may require further analysis to determine various characteristics and behaviors. Consequently, routing logic 235 is configured so as to route information associated with the object 160 to one or more appropriate analysis logic systems, including the static analysis logic system 120₁, dynamic analysis logic system 120₂, and/or the emulation logic system 120₃. A data store 237 may also be used to provide local storage for extraction analysis and rules, as well as operate as a local log. Based on experiential knowledge or machine-learning, the extraction analysis and rules may be periodically or aperiodically updated from a remote source (e.g., threat intelligence network 290).

Referring still to FIG. 2, the static analysis logic system 120₁ is configured to inspect information associated with the object 160 using logic models 240 for anomalies in characteristics such as formatting anomalies for example. In some embodiments, the static analysis logic system 120₁ may also be configured to analyze the object 160 for certain characteristics, which may include the object's name, type, size, path, or protocols. Additionally or in the alternative, the static analysis logic system 120₁ may analyze the object 160 by performing one or more checks, including one or more signature checks, which may involve a comparison between (i) content of the object 160 and (ii) one or more pre-stored signatures associated with known malware. In one embodiment, such information may be stored on the data store 245. Checks may also include an analysis to detect exploitation techniques, such as any malicious obfuscation, using for example, probabilistic, heuristic, and/or machine-learning algorithms.

Additionally, the static analysis logic system 120₁ may feature a plurality of rules that may be stored on the data store 245, for example, wherein the rules control the analysis conducted on the object 160. The rules may be based on experiential knowledge, including but not limited to machine learning; pattern matches; heuristic, probabilistic, or determinative analysis results; analyzed deviations in messaging practices set forth in applicable communication protocols (e.g., HTTP, TCP, etc.); analyzed compliance with certain message formats established for the protocol (e.g., out-of-order commands); and/or analyzed header or payload parameters to determine compliance. It is envisioned that the rules may be updated from an external source, such as via a remote source (e.g., threat intelligence network 290), in a periodic or aperiodic manner.

In one embodiment, the static analysis logic system 120₁ may include vector logic 247 that is configured to determine source information with regard to the object 160. In one embodiment, it is envisioned that the determinations made on behalf of the vector logic 247 may be performed quickly, and thus may logically and often precede other determinations made by the static analysis 120₁. For example, in one embodiment, the vector logic 247 may be configured to determine the source of the object 160, e.g., whether the object 160 was communicated via a communication scheme including, but not limited or restricted to email, file-share, file transfer protocol (FTP), web download, etc., without limitation. In one embodiment, vector logic 247 may be configured to analyze the object 160 to determine source device type or version information associated with the object 160. Thus, it should be understood that the object 160 may be associated with multiple vectors. Accordingly, it is envisioned that the determination with regard to maliciousness may be predicated, at least in part, upon analysis by the vector logic 247.

It is envisioned that information associated with the object 160 may be further analyzed using the dynamic analysis logic system 120₂. Herein, the dynamic analysis logic system 120₂ comprises a virtual machine manager 250, a data store 252 and one or more virtual machines (VMs) 255, namely VM₁ 255₁-VM_M 255_M (M≥1). The VMs 255 are configured to perform in-depth dynamic (behavioral) analysis on the object 160 during processing in order to detect anomalous behaviors. In general terms, the VMs 255 simulate a run-time environment as expected based on the type of object 160. Herein, the dynamic analysis logic system 120₂ is adapted to provision one or more VMs 255₁-255_M (e.g., VM₁-VM_M) using information from the extraction logic 110 (e.g., meta data 161) and/or information from the static analysis logic system 120₁ (e.g., information gathered by the vector logic 247).

Monitoring logic 257 within the dynamic analysis logic system 120₂ may observe one or more behaviors with respect to the object 160 during processing within one or more VMs 255, where these behaviors may be used in a determination by the classification engine 130 as to whether the object 160 is malicious (i.e. a likelihood of maliciousness for the object 160). Using the dynamic analysis logic system 120₂, certain artifacts that may be unique with respect to a type of malware attack may be analyzed. For example, dynamic analysis logic system 120₁ may consider propagation mechanisms of the object 160, to determine how instructions and/or behaviors associated with the object 160 communicate or navigate across and/or through a network, for example.

According to one embodiment, each of the VMs 255 (e.g., VM₁-VM_M) within the dynamic analysis logic system 120₂ may be configured with a software profile corresponding to a software image stored within the data store 252 that is communicatively coupled with the virtual machine manager 250. Alternatively, the VMs 255 (e.g., VM₁-VM_M) may be configured according to a prevalent software configuration, software configuration used by a network device within a particular enterprise network (e.g., client device 220), or an environment that is associated with the object to be processed, including software such as a web browser application, PDF™ reader application, data processing application, or the like.

However, it should be understood for a known vulnerability that the VMs 255 (e.g., VM₁-VM_M) may be more narrowly configured to profiles associated with vulnerable modules. For example, if the access source comprises a certain memory type, VM₁ 255₁-VM_M 255_M may be configured for faster processing and corresponding log file generation. Similarly, when relevant, if the access source is attempting to access a particular application, email address book, etc., then VM₁-VM_M 255 may be configured accordingly.

As further shown in FIG. 2, the object 160 may be further analyzed using the emulation logic system 120₃, which is configured so as to enable the malware detection system 100 (“host” system) to behave like any another computer system (“guest” system). It is envisioned that the emulation logic system 120₃ may be configured so as to enable the host system to run any of various software, applications, versions and the like, designed for the guest system. More specifically, under control of request processing logic 260, the emulation logic system 120₃ may be configured so as to model hardware and software. As such, the emulation logic system 120₃ may be divided into modules that correspond roughly to the emulated computer's various systems, and as a result, the emulation logic system 120₃ comprises subsystems 265. In one embodiment, the subsystems 265 comprise any of various processor emulator/simulators, a memory subsystem module, and/or various I/O devices emulators.

It should be understood that the extraction logic 110, the static analysis logic system 120₁, the dynamic analysis logic system 120₂, the emulation logic system 120₃, the classification engine 130, and/or the reporting engine 150 may be implemented as one or more software modules executed by one or more processors as shown in FIG. 3 below. The processor(s) may be located within the same processor package (e.g., different processor cores) and/or located at remote or even geographically remote locations that are communicatively coupled (e.g., by a dedicated communication link) or a network.

Referring now to FIG. 3, an exemplary embodiment of a logical representation of the malware detection system 100 of FIGS. 1A-1C and 2, deployed as a network appliance 300, is shown. In one embodiment, the network appliance 300 comprises a housing 310, which is made entirely or partially of a rigid material (e.g., hardened plastic, metal, glass, composite or any combination thereof) that protect circuitry within the housing 310, namely one or more processors 320 that are coupled to communication interface logic 330 via a first transmission medium 335. Communication interface logic 330 enables communications with other malware detection systems 100N and/or the threat intelligence network 290 of FIG. 2, for example. According to one embodiment of the disclosure, communication interface logic 330 may be implemented as a physical interface including one or more ports for wired connectors. Additionally, or in the alternative, communication interface logic 330 may be implemented with one or more radio units for supporting wireless communications with other electronic devices.

Processor(s) 320 may further be coupled to persistent storage 340 via a second transmission medium 345. According to one embodiment of the disclosure, persistent storage 340 may include one or more software modules associated with the functionality of the malware detection system 100, which may include the extraction logic 110, the static analysis logic system 120₁, the dynamic analysis logic system 120₂, the emulation logic system 120₃, the classification engine 130, and/or reporting engine 150. It is envisioned that one or more of these systems (or engine) could be implemented externally from the malware detection system 100 without extending beyond the spirit and scope of the present disclosure.

Although not illustrated, it is contemplated that all or a portion of the functionality of the malware detection system 100 may be deployed at part of cloud services. For instance, dynamic analysis logic system 120₂ of FIG. 2 may be deployed in cloud services, which conducts behavioral analysis on an object. Alternatively, at least some functionality of the classification engine 130 of FIG. 2 may be deployed within cloud services. Alternatively, some or all of the functionality of the malware detection system 100 may be deployed as part of client device functionality, where a security agent deployed in the client device performs such functionality described herein. Therefore, it is within the spirit of the invention that the malware detection system may reside entirely within a single network appliance or device, or may be deployed as a decentralized system with different functionality being handled by different network devices, including cloud services.

B. General Analysis Methodology with Respect to the Plurality of Analyses

Referring now to FIG. 4, a flowchart of an exemplary method for analyzing an object using a plurality of analyses is shown. Herein, the malicious detection system receives an object (block 400). Thereafter, the received object is analyzed using any of the various analyses as discussed herein such as static analysis, dynamic analysis, and/or emulation analysis (block 405). An initial determination may be made with respect to maliciousness of the object if there is any attributes (e.g., characteristic(s), behavior(s), and/or emulated result(s)) that present dispositive evidence thereto (block 410) and whether contextual analysis is to be conducted (block 415). At block 420, an alert may be generated if maliciousness is determined and no further contextual analysis is desired.

Otherwise, as described herein, a contextual analysis is conducted where, in accordance with the expanded correlation rule set, attributes from the different analysis logic systems are correlated to produce one or more multi-type attribute combinations (e.g., one or more characteristics from the static analysis logic system and at least one or more behaviors from the dynamic analysis logic system). The multi-type attribute combinations may correspond to multi-type attribute patterns set forth in the expanded correlation rule set. Thereafter, the multi-type attribute combinations are assigned a threat index value that may be used, at least in part, by the object classification logic when classifying the object as malicious or non-malicious (block 425). Thereafter, a determination may be made with respect to maliciousness, in view of the contextual information that includes one or more multi-type attribute combinations (block 430).

If not, one or more secondary contextual analyses may be conducted in which one or more additional analyses are conducted on the object (or a portion of the object) to confirm a presence (or absence) of one or more particular attributes (feedback 435). Based on one or more secondary contextual analyses to improve accuracy of malware detection, a determination is made as to whether the object is malicious or non-malicious (block 440). If non-malicious, the analysis is stopped for the object (block 445). If the object is classified as malicious, an alert may be generated (block 450).

To provide more detail with respect to the analysis methodology, referring back to FIG. 2, analytic results from each of the various analysis logic systems 120₁-120_N may be routed to the classification engine 130 for further processing. As shown, the classification engine 130 comprises the attribute correlation logic 140, threat index generation logic 142 and the object classification logic 144. According to one embodiment of the disclosure, the object classification logic 144 is configured to classify the object 160 in response to operations by the correlation logic 270. The correlation logic 270 includes the attribute correlation logic 140 that is responsible for correlating attributes 170₁-170₃ in accordance with the extended correlation rule set 136 stored in data store 278. The extended correlation rule set 136 includes attribute patterns that are considered in determining whether or not the object 160 is malicious. Herein, one or more of the attribute patterns are directed a multi-type attribute combination, which may include one or more characteristics and one or more behaviors.

The correlation logic 270 further includes threat index generation logic 142 that includes an index assignment logic 272 and a weighting logic 274. Responsive to detecting contextual information (correlated attributes) that correspond to a particular attribute pattern, the index assignment logic 272 assigns a corresponding threat index value to the contextual information while the weighting logic 274 may apply a weighting (when applicable). It is contemplated that the application of a threat index value and/or weighting may be applied to each attribute (i.e. each characteristic [C], behavior [B], or emulated result [E]) or may be applied to each combination of attributes, notably multi-type attribute combinations (e.g., [C, B], [C, E], or [B, E]).

Previously, the analytic results from the analysis logic systems (static analysis logic system 120₁, dynamic analysis logic system 120₂, and/or emulation logic system 120₃) have been considered independently of each other to determine that a particular object 160 is non-malicious. However, to provide greater accuracy with regard to such determinations, the attribute correlation logic 140 within the correlation logic 270 is configured to correlate at least one or more characteristics and one or more behaviors as resultant contextual information, where the detected presence of the contextual information is assigned an threat index value that influences a determination as to whether the object is malicious or not and decrease the incidence of “false negative” and/or “false positive” outcomes.

By way of illustrative example, consider an organization that receives an object 160 comprising a file with a “. SCR” extension that features hidden malware. Typically, “. SCR” files are executable files that can be configured to display vector graphic or text animations, play slide shows, animation, etc., and may include sound effects. In other words, the “.SCR” file is typically associated with benign, non-malicious behavior. Thus, when considered independently (for example, using the static analysis logic system 120₁), a determination may be erroneously made that the .SCR file is non-malicious (benign), demonstrating the potential for returning “false negative” outcomes.

Accordingly, to prevent such “false negative” outcomes and harm to a client's device, the correlation logic 270 is configured to consider the object 160 in view of the other analyses. For example, when the .SCR file is examined by the dynamic analysis logic system 120₂ and/or the emulation logic system 120₃, certain behaviors and emulated results may be determined, respectively that indicate that the .SCR file is actually malicious. More specifically, the dynamic analysis logic system 120₂ may reveal that when executed, the .SCR file actually includes instructions to perform network activity (such as performing an HTTP ‘GET’ request or any other similar instruction), which may be consistent with maliciousness. Accordingly, the classification engine 130 may consider the behaviors as determined from the dynamic analysis logic system 120₂ in view of the characteristics determined from the static analysis logic system 120₁, thereby providing greater accuracy in determining maliciousness.

It is envisioned that based on the file size and/or complexity of the object 160, the various analyses as discussed herein may begin and/or complete at differing times. Thus, it should be appreciated that the various analyses may be performed concurrently (at same time or in an overlapping manner), serially, and/or recursively, without limitation and secondary contextual analyses may be necessary to confirm the presence or absence of certain analytic results.

Using another illustrative example, in one embodiment, the object 160 of FIG. 2 may be observed using the dynamic analysis logic system 120₂, which may reveal behaviors such as network activities that appear to be benign when considered independently of any of the other analysis logic systems. For example, the object 160 may include a link that is configured to navigate to a certain URL (“first destination”) on the Internet. However, prior to indicating that the object 160 is benign or malicious due to this aspect, context logic 276 of the correlation logic 270 may be configured to conduct a “recursive” (or secondary) contextual analysis to consider characteristics determined from the static analysis logic system 120₁ based on various rules, experiential knowledge, signatures and the like that may be stored on the data store 278. Using such information, the context logic 276 may query the static analysis logic system 120₁ (or access locally stored characteristics) to determine by way of non-limiting example, whether the object 160 was signed or un-signed; whether the object 160 was packed or unpacked; and/or the communication protocol of the object 160. Hence, the static analysis logic system 120₁ may reveal, for example, that the object 160 actually included a URL that was supposed to navigate to a second destination, rather than observed first destination. Thus, by noting such inconsistencies and considering the various determinations in such a contextual manner, the incidence of “false negative” outcomes is greatly reduced.

Similar, in another embodiment, the object 160 of FIG. 2 may be observed using the dynamic analysis logic system 120₂, which may include behaviors such as network activities that appear to be malicious when considered independently of any of the other analysis logic systems. However, the correlation logic 270 may correlate the behaviors with certain characteristics that may reveal, for example, that the behavior monitored during processing of the object 160 are actually non-anomalous. Thus, by considering a collection of attributes together as a multi-type attribute communication and in such a contextual manner, the incidence of “false positive” outcomes is greatly reduced as well.

In yet another embodiment, the results of the static analysis may be used to update the dynamic analysis and/or the emulation. In one embodiment, aspects of the object 160 may be analyzed in any order by the static analysis logic system 120₁, the dynamic analysis logic system 120₂, and/or the emulation logic system 120₃. Indeed, in one embodiment, the various analyses as discussed herein may be applied at various times, and to various portions of the object 160.

In other words, any analysis can be revisited, or performed more than once, as needed. Thus, the malware detection system 100 is highly configurable and adaptable depending on the types of objects 160 and corresponding content that is received in order to reduce “false negative” outcomes as well as “false positive” outcomes.

C. General Indexing Methodology with Respect to the Plurality of Analyses

Referring now to FIG. 5, a flowchart of an exemplary method for applying threat index values to observed attributes (e.g., characteristic(s), behavior(s) and/or emulated results) associated with an object under analysis is shown. For example, an object is received and subsequently analyzed using any of the various analyses such as the static analysis, the dynamic analysis, and/or the emulation analysis (block 500).

For example, a static analysis of the object may detect one or more characteristics (C₁, C₂, C₃, . . . C_x; x≥₁), where these attributes (characteristics) are provided to the correlation logic. Similarly, the dynamic (behavioral) analysis of the object during processing within a protective environment (e.g., virtual machine) may observe one or more behaviors (B₁, B₂, B₃, . . . B_y; y≥1), where these attributes (behaviors) also are provided to the correlation logic. As an optional analysis, an operating environment for the object may be emulated and the emulated operations by the object may determine one or more emulated results (E₁, E₂, E₃, . . . E_z; z≥1).

According to one embodiment, these attributes may be correlated by the correlation logic to determine the presence or absence of attribute patterns that are assigned threat index values and are considered in determining whether the object is malicious (block 505). Herein, the attribute patterns may include multi-type attribute patterns that include attributes from different analysis types {[C₁, C₂, B₁]; [C₁, C₃, B₁, B₂]; [C₁-C₄, B₁-B₃, E₂]; [B₁, B₂, E₁]; [C₁, E₁, E₂]; etc.}. It is contemplated that only multi-type attribute patterns exceeding a certain predetermined threat index value may be evaluated to reduce system load.

Upon determining some multi-type attribute combinations {[C₁, C₂, B₁]; [C₁, C₃, B₁, B₂]} corresponding to the multi-type attribute patterns {(C₁, C₂, B₁); (C₁, C₃, B₁, B₂)}, these multi-type attribute combinations may be assigned a corresponding threat index value by the indexing assignment logic (block 510). Of course, it is contemplated that each analytic result may be individually assigned a threat index value ([C₁]; . . . ; [C_x]; [B₁]; . . . ; [B_y]; [E₁]; . . . ; [E_z]).

Optionally, although not shown, certain weights may be assigned to the threat index values using the weighting logic of FIG. 2. In general, the assigned threat index values may be combined and compared with a predetermined threshold value that may vary depending on any of various factors, including by way of non-limiting example, the number and/or existence of various API calls, request-response sessions, and geographical information regarding the most recent, and/or previously targeted parties. Accordingly, it should be understood that the predetermined threshold value might vary for different types of malware attacks. Of course, in other instances, the threshold value may be adapted to accommodate any number of attributes depending on the type of malware attack.

At block 515, according to one embodiment, the threat index for the object may be determined based on the aggregate of threat index values associated with selected attributes and/or combinations of attributes provided from the threat index generation logic. Herein, the threat index may represent a score that represents a likelihood that the object is or contains malware, and in some embodiments, a degree of severity of the attack representative by the selected attribute or combination of attributes. For example, unauthorized exfiltration of data may represented as high severity. However, it is contemplated that the threat index may also be used to “flag” a presence of an object that violates a policy of the customer or identify an ordering of contextual analyses where multiple objects are being examined concurrently. In fact, a secondary contextual analysis may be repeated iteratively until maliciousness can be determined within a certain range of probabilities. Additionally, or in the alternative, the contextual analysis may be completed when all possible combinations of characteristics, behaviors, and/or emulated results have been analyzed and/or appropriately “indexed”. Of course, the analysis could also end after a certain subset of combinations have been considered.

At block 520, an alert may be generated in response to the object being determined to be malicious based on the threat index. More specifically, once the classification engine 130 of FIG. 2 has determined to a degree of certainty (often substantially less than 100%) that the object 160 may be malicious and/or benign, the classified result 175 is communicated to the alert generation logic 280 of the reporting engine 150, which may generate the alert 180 and/or route contextual information that may be part of the alert 180 to security administrators for example, communicating the urgency in handling and preferably preventing one or more predicted attacks. The alert 180 may also include detailed instructions pertaining to specific attack types, potential issues thereto, security holes, and best practices to prevent one or more predicted malware attacks.

It should be appreciated that the reporting engine 150 may also be configured to update the threat intelligence network 290 with information corresponding to the instantly analyzed object 160 for future reference and/or further processing. Additionally, the reporting engine 150 may be configured so as to store the classified results in the data store 285 for future reference. Specifically, the data store 285 may be configured so as to tracking vulnerability exploitation, memory corruption, arbitrary code execution, and other definitive malicious actions, and correlate such data as needed.

D. Exemplary Alert

Referring to FIG. 6, an exemplary embodiment of a mobile network device 600 (e.g., smartphone, tablet, laptop computer, netbook, etc.) configured for performing the contextual analyses described above is shown. Herein, the mobile network device 600 includes a display screen 605 and a housing 610 that encases one or more processors (“processor(s)”) 620, memory 630, as well as one or more receiver and/or transmitter (e.g. transceiver) 640 communicatively coupled to an antenna 645. Herein, the memory 630 includes a security agent 650.

Upon execution by the processor(s) 620, the security agent 650 (executable software application) conducts static and dynamic analyses of at least a portion of information (referred to as “suspect object 660) received by the transceiver 640 of the endpoint device 620. As described above, the security agent 650 conducts a static analysis of the suspect object 660 to obtain certain characteristics and conducts a dynamic analysis of the suspect object 660 in accordance with the fully-instrumented software profile. The characteristics and certain monitored behaviors by the suspect object 660 during processing may be stored in an event log that is allocated within memory 630.

Herein, memory 630 may further include the classification engine 670 and reporting engine 680 that operate in combination with the security agent in a manner similar to classification engine 130 and reporting engine 150 of FIGS. 1A-1C and 2 as described above. In particular, multi-type attribute patterns are evaluated by correlation logic within the classification 670 and detected multi-type attribute combinations are assigned index values and are considered in classifying the suspect object 660 as malicious or not.

In one embodiment, an exemplary alert 690 (e.g., an object, text message, display screen image, etc.) is communicated to security administrators and/or may be displayed for viewing on the mobile network device 600. For example, the exemplary alert 690 may indicate the urgency in handling one or more predicted attacks based on the maliciousness of the suspect object. Furthermore, the exemplary alert 690 may comprise instructions so as to prevent one or more predicted malware attacks. The exemplary alert 690 may also comprise information with respect to the origination of the potential attack, along with suspicious behavior that might confirm the attack with respect to a potential target. In one embodiment, the exemplary alert 690 may include index values represented as scores and/or percentages based on the various analyses and/or the combination of detected behaviors, characteristics and/or emulated results that caused the alert 690, as discussed herein.

In the foregoing description, the invention is described with reference to specific exemplary embodiments thereof. However, it will be evident that various modifications and changes may be made thereto without departing from the broader spirit and scope of the invention as set forth in the appended claims.