Intelligent context aware user interaction for malware detection

Abstract

According to one embodiment, a malware detection system is integrated with at least a static analysis engine and a dynamic analysis engine. The static analysis engine is configured to automatically determine an object type of a received object. The dynamic analysis engine is configured to automatically launch the object after selecting an action profile based on the object type. The dynamic analysis engine is further configured to, provide simulated user interaction to the object based on the selected action profile either in response to detecting a request for human interaction or as a result of a lapse of time since a previous simulated human interaction was provided.

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 for detecting malware.

GENERAL BACKGROUND

Over the last decade, malicious software (malware) has become a pervasive problem for Internet users. Often malware exploits vulnerabilities in networked resources. For instance, over the past few years, more and more 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 exploits that use malicious computer code. The malware may attempt to acquire sensitive information or adversely influence or attack normal operations of a network device or the entire enterprise network.

Currently, in malware detection systems, one or more virtual machines may be used to process objects, which may include, for example, content from network traffic and/or files retrieved from a storage location, in order to activate, observe, and thereby detect malicious software. However, this processing may require user interaction, for example, in the form of an input initiated by an input device such as a graphical user interface (GUI), mouse, keyboard, keypad or the like. Based on an inability to provide the necessary user input, current malware detection systems may fail to activate the malicious content within the objects. One reason is that sophisticated malware often has a self-defense mechanism, which attempts to detect whether it is running in a virtual environment of a malware detection system rather than the intended environment of a client device under user control. One type of self-defense mechanism involves the malware monitoring whether user input expected by an application is supplied at the appropriate time. If it is not, the malware may simply hibernate (not activate), and thus not present itself for detection by the malware detection system.

Some conventional malware detection systems apply generic, static patterns of simulated input device controls in a virtual run-time environment in the absence of actual human interaction. However, malware creators have been able to identify these patterns. As a result, they have been able to equip their malware to identify such static simulated device controls, and upon detection, cause the malware to refrain from activating the malicious code in order to remain undetected. As a consequence, some conventional malware detection systems may experience unacceptable levels of false negatives or be forced to deploy a multitude of pattern detection schemes that will increase the rate of false positives.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the disclosure 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. 1 is an exemplary block diagram of a first embodiment of a malware detection system (MDS) which features user interaction (UI) control logic deployed as part of a virtual run-time environment to provide simulated user interaction to detonate a malicious object.

FIG. 2 is an exemplary block diagram of a second embodiment of the MDS which features a second architecture scheme of the UI control logic deployed as part of the virtual run-time environment.

FIG. 3A is an exemplary block diagram of a first operational flow for the UI control logic operating within the virtual run-time environment.

FIG. 3B is an exemplary block diagram of a second operational flow for the UI control logic operating within the virtual run-time environment.

FIG. 4A is a flowchart illustrating an exemplary process conducted by passive UI simulation logic and device control simulation logic that are operating in accordance with a selected action profile and are conducting simulated user interactions to increase the likelihood in accurately detecting a malicious object.

FIG. 4B is a flowchart illustrating an exemplary process conducted by the active UI simulation logic that is operating in accordance with the selected action profile and is conducting simulated user interactions to increase the likelihood in accurately detecting a malicious object.

FIG. 5 is an exemplary block diagram of the logical architecture associated with the MDS of FIG. 1.

DETAILED DESCRIPTION

Various embodiments of the disclosure are directed to a system and method for determining whether or not an object is associated with a malicious attack through, at least, a dynamic analysis of the object within a virtual run-time environment. Herein, the virtual run-time environment features one or more virtual machine instances (VMs), which are provisioned with a guest image associated with a prescribed software profile. The guest image may include a software application in addition to an operating system (OS) along with monitors, namely software components that are configured to observe and capture run-time behavior of an object under analysis during processing within the virtual machine. With this VM configuration, in order to effectively detect a malicious object, the object needs to be “launched” and subsequently “detonated” within the virtual run-time. Herein, the term “launch” (and other tenses) represents performance of one or more events that starts activation of an object under analysis while the phrase “detonate” (and other tenses) represents performance of one or more events that trigger a malicious attack by the activated object. Additionally, the contents of related application, U.S. patent application Ser. No. 13/801,532 filed on Mar. 13, 2013 is incorporated by reference herein.

In some cases, however, objects are detonated only in response to some sort of user interaction (e.g., one or more user inputs responsive to an event actuated by the object or user-initiated inputs during normal use of the object, etc.). According to one embodiment of the disclosure, user interaction (UI) control logic may be deployed as of the virtual run-time environment in order to provide simulated user interaction needed to detonate certain types of malicious objects within a VM. Embodiments of the invention provide simulated user interaction tailored to the type of object (and, in some embodiments, other features related to the object) being processed in the virtual run-time environment. The UI control logic comprises a plurality of components, including (1) a profile selector and (2) a UI framework, as described herein.

Herein, the UI control logic may feature multiple implementations. For instance, the UI control logic may be provisioned as components of a VM. As an alternative embodiment, the UI framework may be provisioned as a component of the VM, but the profile selector may be deployed as part of a virtual machine monitor (VMM), which may be deployed, according to one embodiment of the disclosure, as part of a “hosted hypervisor” (e.g., software that runs on top of a host operating system) or as an intermediary operation layer between the hardware and the VMs. When deployed as part of the VMM, the profile selector may be adapted to provision the UI framework component within the VM and perhaps multiple UI framework components within multiple VMs.

According to one embodiment of the disclosure, the profile selector selects an action profile from a plurality of action profiles that may be hosted in the VM or outside the VM within the virtual run-time environment. This selection may be based, at least in part, on metadata associated with an object under analysis. Herein, the metadata defines, at least in part, the context for determining the action profile that governs the simulated user interaction. The metadata further determines the software appropriate to launch the object in the run-time environment. Of course, the selection of the action profile also may be based on a type of network device deploying (hosting) the VM (e.g., security appliance that analyzes network traffic, files within a file storage system, etc.) or other information from the static analysis of the object. According to this action profile selection scheme, the dynamic analysis of the object is “context aware”.

Herein, the metadata may include data that identifies the type of object under analysis. Of course, it is contemplated that, besides object type, other metadata may be used by the profile selector for selecting the particular action profile. Examples of other metadata that may be used by the profile selector to select a particular action profile for controlling the simulation of user interactions with the object launched in the VM may include, but are not limited or restricted to information related to the following: (i) whether the object is encrypted and/or its type of encryption scheme, (ii) whether the object is an embedded object, (iii) the type of application needed for processing the object, and/or (iv) transmission protocol used in delivery of network content including the object.

Each “action profile” is a collection of instructions and/or commands that performs UI functionality in accordance with a set of rules prescribed for that action profile. As a result, each action profile is configured for use in dynamically controlling UI actions associated with a certain type of object in contrast to the use of patterns per se. For instance, the action profile associated with a Microsoft® Excel® spreadsheet may conduct different UI actions (e.g., select tabs, add text to certain cells, scroll down a certain number of cell rows, etc.) than a PDF document (e.g., scroll pages of the document, etc.) and such actions may be conducted at different times depending on the behavior of the object under analysis.

As described herein, the UI framework comprises (i) the actuation logic, (ii) active UI simulation logic; (iii) passive UI simulation logic; and (iv) device control simulation logic. According to one embodiment of the disclosure, the actuation logic is software component that is implemented as part of a software profile that provisions the VM and is responsible for launching the object under analysis. The particular implementation of the actuation logic may vary depending on the object type. Upon the actuation logic launching the object, the active UI simulation logic, the passive UI simulation logic and the device control simulation logic are instantiated with or are instantiated to access content within the selected action profile. Operating in accordance with the selected action profile, the simulation logic within the UI framework conducts particular actions (e.g., expected user interface interactions and/or methods of activation) during particular operating states at which such actions are expected if the object was running on a targeted endpoint. These particular actions may be conducted in with a predetermined sequence (order) and/or at (or within) predetermined periods of time. Furthermore, two or more of these particular actions may be conducted concurrently (at least partially overlapping at the same time) or such actions may be performed sequentially.

Operating as part of the UI framework, the active UI simulation logic detects input requests (e.g., password request, opening of a dialog box that requires dismissal prior to continuing, opening of a text box that requires text entry, etc.), which require human interaction that directly responds to the input request. This type of simulated human interaction is referred to herein as “active” simulated human interaction. In response, the active UI simulation logic operates in accordance with the selected action profile to determine whether to provide a response and the type of response, where appropriate.

The passive UI simulation logic operates in accordance with the selected action profile and, in certain cases, provides simulated human interaction in response to a prescribed level of inactivity by the object and/or a prescribed period of time after the suspect object has launched has elapsed. The passive UI simulation logic is in communication with timing circuitry (e.g., real time clock, counter, etc.), where the monitored time plays a factor in determining when to conduct prescribed simulated human interactions that are triggered by a period of inactivity by the object and/or an elapsed time from when the suspect object was launched.

Responsive to detecting a prescribed period of inactivity for example, the passive UI simulation logic simulates user-initiated interactions on the object such as moving to particular page in a Microsoft® Office Word document (object), switching to a tab in a Microsoft® Office Excel document (object), or switching to a different PowerPoint™ slide in accordance with the object-specific action profile. As an assuming the object is a Microsoft® Office Excel document, experiential knowledge of typical placement of exploit/malicious code (e.g., through machine learning techniques) in a Microsoft® Office Excel document may result in instructions by the selected action profile for the passive UI simulation logic to simulate human interaction by switching to the second sheet of the Microsoft® Office Excel document at a predetermined time after the actuation logic launches the object.

The device control simulation logic operates in accordance with the selected action profile and provides simulated device controls that are agnostic to object type, which may occur in response to yet another level of prescribed inactivity. For example, the device control simulation logic may receive instructions from the selected action profile to simulate certain device control interactions, such as simulate particular keystrokes and/or particular mouse movements, in an attempt to trigger a malicious attack by the object.

Embodiments of the disclosure may be employed by or take the form of a network device, including a cyber-security appliance that features a malware detection system (MDS). The MDS includes a static analysis engine and a dynamic analysis engine, or, in another embodiment, only a dynamic analysis engine. In some embodiments, the MDS may be implemented as a server or client device or other system (any of which may be referred to as an “endpoint”) connectable to a network. The dynamic analysis engine may include a virtual run-time environment that automatically analyzes, without user assistance, objects from the received network traffic and simulates human interaction to detonate and detect malicious objects during virtual processing. The results of the analysis may be reported to network administrators or other personnel for further analysis and action.

I. Terminology

In the following description, certain terminology is used to describe features of the invention. For example, in certain situations, the terms “logic”, “component”, and “engine” are representative of hardware, firmware and/or software that is configured to perform one or more functions. As hardware, logic (or component or engine) 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 processors and/or processor cores, a programmable gate array, a microcontroller, an application specific integrated circuit, semiconductor memory, or combinatorial logic.

Logic (or component or engine) 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 may be stored in persistent storage.

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 the presence of malware and potentially allow the object to be classified as part of a malicious attack.

Examples of objects may include one or more flows or a self-contained element within a flow itself. 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, according to one embodiment, include packets, frames, or cells. Further, an “object” also refer to collective payloads of a number of related packets, e.g., a single webpage received over a network. Moreover, an object may be a file or document retrieved from storage location over a transmission medium.

As a self-contained element, the object may be an executable (e.g., an application, program, segment of code, dynamically link library “DLL”, etc.) or a non-executable. Examples of non-executables may include a document (e.g., a Portable Document Format “PDF” document, Microsoft® Office® document, Microsoft® Excel® spreadsheet, etc.), an electronic mail (email), downloaded web page, or the like.

The term “transmission medium” may be construed as a physical or logical communication path between two or more network devices (e.g., any devices with data processing and network connectivity such as, for example, a security appliance, a server, a mainframe, a computer such as a desktop or laptop, netbook, tablet, firewall, smart phone, router, switch, bridge, etc.) or between components within a network device. For instance, as a physical communication path, wired and/or wireless interconnects in the form of electrical wiring, optical fiber, cable, bus trace, or a wireless channel using infrared, radio frequency (RF), may be used.

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, a security appliance, or the like.

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

The term “action profile” should be interpreted as a plurality of instructions and/or commands that provision logic to conduct, in accordance with a set of rules prescribed for that particular action profile, different types of simulated user interactions. The simulated user interactions may include “active” simulated human interactions; “passive” simulated human interactions and simulated device control interactions.

An active simulated human interaction includes simulated actions that may be performed by a user in response to an event initiated by a suspect object under analysis. In some situations, the simulated action may be required before any further activities are conducted by the object. Examples of an active simulated human interaction include closing a window or dialog box; selecting a particular radio button; and/or entering characters into a text box).

A passive simulated human interaction includes simulated actions that are normally performed by a user during activation of the object, but such actions are not responsive to a particular behavior by the object. Examples of passive simulated human interaction include scrolling pages of a document (e.g., PDF or Word® document), browser, or other type of displayed image; selecting certain tabs of an Excel® spreadsheet; and/or accessing certain menu options.

A simulated device control interaction includes simulated input from an input device for an endpoint. Examples of a simulated device control interaction include keystrokes, mouse movement or clicks, and/or detected activation of certain area or areas of a touch screen.

Lastly, the terms “or” and “and/or” as used herein are to be interpreted as inclusive or meaning any one or any combination. Therefore, “A, B or C” or “A, B and/or C” mean “any of the following: A; B; C; A and B; A and C; B and C; A, B and C.” An exception to this definition will occur only when a combination of elements, functions, steps or acts are in some way inherently mutually exclusive.

The invention may be utilized for detection, verification and/or prioritization of malware, which may include malicious content, in particular, through providing object-type specific simulated human interaction to an object activated in a virtual run-time environment. 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. General Architecture of the Malware Detection System

Referring to FIG. 1, an exemplary block diagram of a network 100 deploying a plurality of malware detection systems (MDS) 110₁-110_N (N>1, where N=3 for this embodiment) communicatively coupled to a management system 120 via a network 125 is shown. In general, the management system 120 is adapted to manage each MDS 110₃. For instance, the management system 120 may be configured to perform content updates (e.g., upload new rules or modified rules, delete rules, modify parameters that utilized by the rules and/or upload metadata) within logic operating as part of a communication interface 140, a static analysis engine 145, a dynamic analysis engine 160, a classification engine 190, and/or a reporting engine 195 with an optional user interface capability (e.g., for purposes of initial system set up and configuration). In particular, the management system 120 is configured to propagate updates to one or action profiles (hereinafter “action profile(s)”) via action profile update logic 174. The action profile(s) 188 are used to control the simulation of human interactions with and/or device controls for the object during analysis, as described below.

As shown in FIG. 1, a first malware detection system (MDS) 110₁ is an electronic device that is adapted to analyze information associated with incoming data (e.g., network traffic, input data over a communication network 105, input data from another type of transmission medium, etc.) from/to one or more endpoints 132. As this illustrative embodiment, the communication network 105 may include 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. Other embodiments may include incoming data (files from a file store) being retrieved from a file storage location (e.g., a hard disk drive (HDD) or a flash drive storage) for malware detection. Herein, the first MDS 110₁ is communicatively coupled with the communication network 105 via an interface 136 and/or an operational firewall 134.

In general, the interface 136 may operate as a data capturing device that is configured to receive at least a portion of network traffic propagating to/from one or more endpoints 132 and provide information associated with the received portion of the network traffic to the first MDS 110₁. This information may include an object, namely multiple packets collectively forming an executable or a non-executable (e.g., a document embedded within an email message or a web page). Alternatively, although not shown, the interface 136 may be configured to receive files or other objects that are not provided over a network. For instance, as an example, the interface 136 may be a data capturing device that automatically (or on command) accesses data stored in a storage system or another type of interface, such as a port, for receiving objects manually provided via a suitable dedicated communication link or from storage media such as portable flash drives.

In general terms, the interface 136 may be configured to capture data typically directed to the endpoint 132, where the captured data includes at least one object 147 for analysis and its corresponding metadata 148. The metadata 148 may be used, at least in part, to determine protocols, application types and other information that may be subsequently used by logic, such as a scheduler 150 for example, to configure one or more VMs 170₁-170_M (M≧1) with selected software profiles. For instance, the metadata 148 may be used to determine which software images (e.g., application(s)), if any, in addition to operating systems to be fetched a storage device 151 for configuring operability of the VMs 170₁-170_M in order to process the subject object 147 at a desired time, for a desired period of time, and/or in a desired order. Additionally, as one feature of the invention, the metadata 148 associated with the suspect object 147 may be used by the profile selector 184, implemented in the VM(s) itself (e.g. VM 170₁) or in the virtual machine monitor (VMM) 172 as described below, to select which action profile(s) 188 for controlling simulated user interaction with the suspect object 147 in efforts to detonate the suspect object 147 within one or more of the VM 170₁-170_M.

In some embodiments, although not shown, interface 136 may be contained within the first MDS 110₁. In other embodiments, the interface 136 can be integrated into an intermediary device in the communication path (e.g., a firewall, router, switch or other networked electronic device) or can be a standalone component, such as an appropriate commercially available network tap as shown.

As further shown in FIG. 1, the first MDS 110₁ comprises the communication interface 140, the static analysis engine 145, the dynamic analysis engine 160, the classification engine 190 and the reporting engine 195. Herein, the communication interface 140 receives an object from the interface 136 and converts that object into a format, as needed or appropriate, on which analysis by the static analysis engine 145 may be conducted. This conversion may involve decompression of the object, decompilation of the object, extraction of specific data associated with the object, and/or emulation of the extracted data (like Javascript™). The communication interface 140, in some embodiments, may be integrated into the interface 136.

Referring still to FIG. 1, the static analysis engine 145 may include one or more controllers (e.g., processing circuitry such as one or more processors) that feature static analysis logic 152, metadata extraction logic 154, and object-type determination logic 156 for accessing magic number database 158 and/or data store 159. The data store 159 may be used to store the analyzed object 147 and/or extracted metadata 148 prior to transmission to the data analysis engine 160 upon determining that the object 147 is “suspicious” and requires further analysis to determine whether that object 147 is associated with a malicious attack.

Referring still to FIG. 1, the static analysis logic 152 includes one or more software modules that, when executed by the controller(s), analyzes characteristics associated with the suspect object 147, which may be a portion of network traffic (or downloaded data) according to this embodiment of the disclosure. Such static analysis may include one or more checks being conducted on the object without its execution. Examples of the checks may include (i) heuristics, which is based on rules or policies as applied to the object 147 and may determine whether one or more portions of the object 147 is associated with anomalous or suspicious characteristics (e.g., a particular URL associated with known exploits, or a particular source or destination address etc.) associated with known exploits; or (ii) determinative rule-based analysis that may include blacklist or whitelist checking.

The metadata extraction logic 154 is responsible for extracting and/or generating metadata 148 contained as part of and/or associated with the suspect object 147. The extraction and/or generation of the metadata 148 may occur after the object 147 is determined by the static analysis logic 152 to feature anomalous or suspicious characteristics. Of course, alternatively, the metadata extraction logic 154 may extract and/or generate the metadata 148 prior to or concurrently with the operations conducted by static analysis logic 152.

The metadata 148 may be identified as being associated with the suspect object 147, and is stored accordingly. Examples of metadata 148 may include, but are not restricted or limited to, information that identifies the type of object 147. For example, a particular document (e.g., Microsoft® Excel spreadsheet) is an example of an object type, which may be in the form of a non-executable. This metadata 148 may be subsequently used by the profile selector 184 to select at least one particular action profile for controlling simulated user interaction conducted during analysis of the object 147 within one or more VMs 170₁-170_M of the virtual run-time environment 164, as described below.

In addition to, or in lieu of the metadata associated with the source of the object 147, it is contemplated that other metadata may be captured by metadata extraction logic 154. For instance, other metadata which may be used by the profile selector 184 for selecting the particular action profile. This metadata may include, but is not limited or restricted to the following: (i) data identifying whether the object is encrypted and/or its type of encryption scheme, (ii) data identifying whether the object is an embedded object, (iii) data identifying the type of application needed for processing the object, and/or (iv) data identifying the transmission protocol used in delivery of network content including the object. These and potentially other features related to the object are stored for later use.

From the extracted metadata, the object-type determination logic 156 may determine object type. For instance, the object-type determination logic 156 may analyze content within the object 147, which may identify the object type. For instance, as an illustrative example, the object-type determination logic 156 may identify a predetermined number of bytes at the beginning of the object 147 (sometimes referred to as the “magic numbers” for the object) and compare the values associated with these bytes with stored values within the magic number database 158. Upon a successful comparison, the object-type determination logic 156 has identified the object type.

For instance, as an illustrative embodiment, the first few bytes of the object 147 may, in certain cases, be used to determine the object-type or at least infer the object based on the communication protocol in use. As an example, the object-type determination logic 156 may determine that the object 147 starts with the hexadecimal string value “4D5A” which, upon comparison with entries within the magic number database 158, identifies that the object 147 is an executable. Similar, the object-type determination logic 156 may determine that the object 147 starts with a hexadecimal string value of “25 50 44 46” and, upon comparing this value with stored data within the magic number database 158, determines that the object 147 is a PDF document.

As discussed above, the static analysis engine 145 may route the suspect object 147 along with the metadata 148 (inclusive of any object type information generated by the object-type determination logic 156) to the virtual run-time environment 164 within the dynamic analysis engine 160. The results of the static analysis may be used to establish an order of processing of objects in the virtual run-time environment 164 based on the level of “suspiciousness” of the objects (e.g., as established by a relative suspiciousness score). The static analysis engine 145 may also filter benign objects from further analysis. In one embodiment, if the object 147 does not appear suspicious and/or malicious based on a static analysis, the static analysis engine 145 may simply denote that the object 147 is non-malicious and may refrain from subjecting the object 147 to further analysis. However, upon determining that the object 147 includes characteristics that are suspicious, extracting the metadata 148 associated with the suspect object 147 and determining the object type, the static analysis engine 145 may pass this suspect object 147 along with the metadata 148 to the dynamic analysis engine 160 for more in-depth analysis in a VM-based operating environment. All or portions of the static analysis engine 145 may be integrated into the interface 136 or into other devices, such as a firewall or another network device, such as a network device located at the periphery of a network to be protected so as to capture and examine objects contained in ingress content.

The dynamic analysis engine 160 may include processing logic 162, a virtual time environment 164, a data store 166, and/or a score determination logic 168. According to one embodiment, processing logic 162 may be configured to control inter-operability between components within the dynamic analysis engine 160. For instance, the processing logic 162 may control the buffering of the passed objects and their corresponding metadata into the data store 166 and the loading of the objects and corresponding metadata into the VM(s) 170₁-170_M directly or into the VMM 172 for supply to the VMs 170₁-170_M.

The virtual run-time environment 164 provides for virtual processing of the object 147 through one or more VMs 170₁-170_M managed by a virtual machine monitor (VMM) 172. The VMM 172 manages reconfiguration of the one or more VMs 170₁-170_M before conducting the virtual analysis based on externally provided configuration updates, namely software profiles (e.g., OS and/or application instances), action profiles, or the like. As shown, the VMM 172 features action profile update logic 174, which is responsible for updating rules, parameters, instructions, and/or other data maintained by the action profile(s) 188 hosted in VM 170₁, as shown. Of course, the action profile update logic 174 may update action profile(s) 188 hosted in other VMs (e.g., VM 170_M) or hosted outside the VM (e.g., within storage device 151, within data store 166, or within storage within the virtual run-time environment (not shown)). For clarity sake, the operations of VM 170₁ are described, although all or some of the other VMs 170_M or VMs 170₂-170_M may operate in a similar manner.

As shown, the VM 170₁ may be provisioned with an operation system (OS) and, dependent on the object type, one or more applications 180, along with the monitoring logic 181 and user interaction (UI) control logic 182. The monitoring logic 181 monitors run-time behaviors of the object 147 when launched in the VM 170₁. The UI control logic 182 provides simulated user interactions to detonate a malicious object that is loaded into the VM 170₁ and requires some sort of user interaction to initiate a malicious attack. According to one embodiment of the disclosure, the UI control logic 182 comprises a plurality of components, which include (1) a profile selector 184 and (2) UI framework 186.

According to one embodiment of the disclosure, the profile selector 184 selects an action profile from the action profile(s) 188 that are shown as being hosted in the VM 170₁. This selection may be based, at least in part, on the metadata 148 associated with the suspect object 147. For example, the metadata 148 may include data produced by the object-type determination logic 156 that identifies an object type for the object 147. As described above, the metadata 148 may include other data that is uncovered during parsing of the object 147 by the static analysis engine 145 (e.g., password protected fields, password in an email message that included the object 147, etc.), which may be relied upon for selecting a particular action profile within the action profile(s) 188.

As further shown in FIG. 3A and FIG. 3B, the UI framework 186 comprises (1) the actuation logic 340 and (2) different types of user interaction (UI) simulation logic 350, where the different UI simulation logic implemented within the UI framework 186 as described below. Herein, the actuation logic 340 is a software component which is implemented as part of the software profile that provisions the VM 170₁ and is responsible for “launching” the suspect object 147. The object may be “launched” by starting a process that starts the object 147 (when the object 147 is an executable or script) or starting a process that manipulates the object 147 (e.g., opens the object 147 using Adobe® Acrobat® Reader application when the object 147 is a PDF document; opens the object 147 using Microsoft® Office® application when the object 147 is an Excel® (.xls) file, etc.).

When launching the object 147, the actuation logic 340 notifies the UI simulation logic 350 of the launched object. In response, logic within the simulation logic 350 is instantiated with or is instantiated to access the selected action profile, which controls simulated user interaction conducted by the UI framework 186 during analysis of the object 147. The simulated user interaction may include signaling that simulates a particular action during a particular operating state of the object 147 at which such an action is expected if running on a targeted endpoint (client device). These particular actions may be order dependent (sequenced) and/or time dependent (e.g., occur at a particular time, occur at a particular time after a previous action, etc.).

Referring back to FIG. 1, the monitoring logic 181 and UI framework log 176 collectively operate to record, while the object 147 is launched in the VM 170₁, the requests for input by the object 147. The monitoring logic 181 and the UI framework log 176 also monitor and log any “active” simulated human interactions provided to the suspect object 147 in response to the input request, any responses to the simulated human interactions by the suspect object 147, “passive” simulated human interactions that are not responsive to behaviors by the suspect object 147, and/or simulated device control interactions. The data recorded by the UI framework log 176 may be referenced by the score determination logic 168, which determines a probability (score) that is used, at least in part by the classification engine 190, to determine (i) whether the suspect object 147 is associated with a malicious attack and (ii) severity of the malicious attack. The contents of the UI framework log 176 may also provide an ability to correlate, for reporting purposes, malicious objects that are detonated in response to certain type(s) of user interaction.

For instance, when a submitted object 147 is classified as malicious, the UI framework log 176 can provide information for understanding which simulation logic caused or helped a successful detonation. In other words, from data within the UI framework logic 176, a determination can be made as to the efficacy of action profiles and the UI framework. Such feedback can be used to “fine-tune” action profiles. Additionally, by use of data within the UI framework logic, malwares can be classified based on user interaction(s) necessary for detonation. This classification and details of user interaction(s) can augment the Threat Intelligence aspects such as forensic analysis of malwares and incidence response. Similarly, when the object 147 is classified as suspicious, the UI framework logic 176 provides information for understanding the shortcomings in the set of user interactions the UI framework 186 provides (e.g., a new feature might be required in UI framework 186 or new rules or parameters may be needed for the selected action profile). On the other hand, if a user interaction performed by the UI framework obstructs object detonation, it can be rectified in subsequent action profile update.

As shown in FIG. 1, the reporting engine 195 is adapted to receive information from the classification engine 190 via transmission medium 189 and generate alerts (e.g., various types of messages including text messages and email messages, display images, or other types of information over a wired or wireless transmission medium) that identify to a network administrator that the suspect object 147 is associated with a malicious attack and is user-interaction dependent.

Although FIG. 1 illustrates the MDS 110₁ as a dedicated network device and the discussion of FIG. 1 explains examples based on an object received over the network interface 136, the MDS 110₁ may be implemented on an endpoint, such as the client device 132. In such an embodiment, prior to actual execution of the object, the MDS 110₁ may launch the object in a sandboxed environment and conduct simulated user interactions, which may include simulated human interactions and simulated device controls. Responsive to non-anomalous behaviors by the object 147, the endpoint 132 is allowed to utilize the object. In addition, the MDS 110₁ may be implemented in the cloud computing services 138, where the above described simulated human and device control interactions may be fully or partially conducted therein.

Referring now to FIG. 2, a block diagram of a second embodiment of the MDS 110₁ employed within a network 200 is shown. According to this embodiment of the disclosure, the dynamic analysis engine 160 includes processing logic 162, virtual run-time environment 164, data store 166 and/or score determination logic 168. However, the VM(s) 170₁-170_M are not provisioned with action profile(s) 188 as shown in FIG. 1. Rather, action profile(s) 188 are hosted outside the VM(s) 170₁-170_M. Furthermore, the profile selector 184 is deployed as part of the VMM 172. For this implementation, the profile selector 184 may be adapted to provision the UI framework component 186 within the VM 170₁ and perhaps UI framework components within other VMs (e.g., VM 170_M).

III. Operation Flow of the UI Control Logic

Referring to FIGS. 3A and 3B, exemplary block diagrams of operational flows for the UI control logic 182 within the virtual run-time environment 164 is shown. In particular, one embodiment of the UI control logic 182 operating within the VM 170₁ is illustrated. Herein, the UI control logic 182 comprises (1) the profile selector 184 and (2) the UI framework 186. As shown, the UI framework 186 comprises (a) actuation logic 340 and (b) simulation logic 350, which includes (i) active UI simulation logic 360; (ii) passive UI simulation logic 370; and (iii) device control simulation logic 380.

According to one embodiment of the disclosure, the object 147 and metadata 148 are provided to the VM 170₁. Based on the metadata 148, the profile selector 184 selects an action profile (herein the “selected action profile” 300₁) within the action profile(s) 188, namely a plurality of action profiles 300₁-300_R (R≧2) that may be hosted in the VM 170₁ (as shown) or outside the VM 170₁ within the virtual run-time environment. This selection may be based, at least in part, on metadata identifying the object type. Of course, it is contemplated that, besides object type, other metadata may be used by the profile selector 184 to better identify the object 147 in order to choose the selected profile 300₁ best suited for the particular object under analysis. Examples of other metadata that may be used include, but are not limited or restricted to the following: (i) data identifying whether the object 147 is encrypted and/or its type of encryption (ii) data identifying whether the object 147 is or contains an embedded object, (iii) data identifying whether the object 147 includes password-protected fields and information associated with the password; (iv) data identifying the type of application needed for processing the object 147, and/or (v) data identifying the transmission protocol used in delivery of network content including the object 147.

Herein, according to one embodiment of the disclosure, each “action profile” is a collection of instructions and/or commands that performs UI functionality in accordance with a set of rules prescribed for that action profile. As a result, the selected action profile 300₁ is configured for use in controlling UI functionality during analysis of the object 147. For instance, where the object 147 is identified as a Microsoft® Excel® spreadsheet, the selected action profile 300₁ may conduct different UI functions (e.g., select tabs, add text to certain cells, scroll down a certain number of cell rows, etc.) than another action profile 300_R for controlling UI functionality during analysis of a PDF document (e.g., scroll pages of the document, etc.).

According to a first embodiment, as shown in FIG. 3A, upon selection of the action profile, the profile selector 184 provides signaling 310 to identify the selected action profile 300₁ that is part of the pre-stored action profile(s) 188. In response, according to one embodiment of the disclosure, the content 320 of the selected action profile 300₁ may be passed to the simulation logic 350 for use by the active UI simulation logic 360, the passive UI simulation logic 370, and the device control simulation logic 380. According to another embodiment of the disclosure, as shown in FIG. 3B, upon choosing the selected action profile 300₁, the profile selector 184 passes an identifier 330 of the selected action profile 300₁ to the simulation logic 350 (active UI simulation logic 360, passive UI simulation logic 370, and device control simulation logic 380) to allow the simulation logic 350 to poll and retrieve information 335 (e.g., commands, instructions, rules and/or parameters) from the selected action profile 300₁.

As an optional feature, although not shown, addressing information (e.g., a pointer, memory storage location, etc.) may be provided to the actuation logic 340 associated with that particular object type. The addressing information may be used for accessing a sequence of commands and/or instructions that conducting operations suitable for launching a particular object type.

Referring to both FIGS. 3A and 3B, implemented as part of a software profile that provisions the VM 170₁, the actuation logic 340 is responsible for launching the object 147. The particular implementation of the actuation logic 340 may vary depending on the object type. For instance, where the object 147 is a document type (e.g., Microsoft® Word® document, PDF document, etc.), the actuation logic 340 may be customized logic which supports launching (in this case, opening) of the object 147 (or concurrent launching of the object 147 by different versions of) the application and/or OS (e.g., Windows® 7 and Office® 2013; Windows® 7, Office® 2010, etc.). Alternatively, where the object 147 is an executable, the actuation logic 340 may be a software module (e.g., script, etc.) that copies the object 147 to a file system storage location and subsequently calls an operating system (OS) function, such as “CreateProcess( )” for example, to process the object 147.

Upon launching the object 147, the actuation logic 340 provides a launch notification 345 to the simulation logic 350, namely the active UI simulation logic 360, the passive UI simulation logic 370 and the device control simulation logic 380. According to one embodiment, the launch notification 345 may cause the simulation logic 350 to poll for data 347. According to one embodiment of the disclosure, the data 347 may include (i) an identifier for the object 147; (ii) an identifier as to a type of actuation logic (e.g., particular software module) used to launch the object 147; and/or (iii) the time that the object 147 was launched. Of course, in accordance with a “push” communication scheme, the data 347 may be provided as part of the launch notification 345.

According to this embodiment, the identifier of the object 147 and/or the identifier of the actuation logic 340 may be used to verify that the correct selected action profile 300₁ has been passed to the simulation logic 350 for use at the correct time(s) during processing of the object 147. The launch time may be used to synchronize the active UI simulation logic 360, the passive UI simulation logic 370 and the device control simulation logic 380 with each other. The launch time also establishes a reference time for use when the passive UI simulation logic 370 is conducting time-simulated human, and/or the device control simulation logic 380 is conducting time-based simulated device control interaction in accordance with the selected action profile 300₁. The synchronization is especially relevant for actions conducted by the passive UI simulation logic 370 and the device control simulation logic 380 in accordance with the selected action profile 300₁, as illustrated in FIG. 4A, and adds intelligence and predictability to the user interaction. Therefore, it is contemplated that most or all of the UI simulation logic 360/370/380 is in communication with the same (or synchronized) timing circuitry (e.g., real time clock, counter, etc.).

As further shown in FIG. 3A and FIG. 3B, the active UI simulation logic 360, the passive UI simulation logic 370 and the device control simulation logic 380 are instantiated with or are instantiated to access content within the selected action profile 300₁, which controls the simulated human and device control interactions conducted by the UI framework 186 during analysis of the object 147. Collectively, in accordance with the rules outlined in the selected action profile 300₁, the simulation logic 350 conducts particular actions (e.g., expected user interface interactions and/or methods of activation) during particular operating states at which such actions are expected (e.g., in predetermined sequence (order) and/or at or within a predetermined period of time). Furthermore, although described as being conducted in sequentially, it is contemplated that two or more actions may be conducted concurrently (at least partially overlapping at the same time).

Operating as part the UI framework 186, the active UI simulation logic 360 is a first type of simulated user interaction which is configured to detect input requests (e.g., password request, an attempt to display a dialog or text box for selection of a radio button or text input, etc.) initiated by the object 147 that require “active” human interaction. In response, based on the contents of the selected action profile 300₁, the active UI simulation logic 360 determines whether to provide a response and, where appropriate, the type of response that simulates the requested human interaction. For instance, the selected action profile 300₁ may cause the active UI simulation logic 360 to provide signaling that simulates human interaction responsive to the input request initiated by the launched object 147. For example, the signaling may simulate the user closing a dialog box that requires dismissal before continuing or simulate the user selecting a particular radio button that closes the dialog box and opens another dialog box for handling. Such signaling may be intentionally delayed by a prescribed or period of time to further simulate human interaction. This response and/or responses to subsequent input requests may trigger the object 147 to commence a malicious attack, which could only have been activated by such simulated human interactions.

The passive UI simulation logic 370 is a second type of simulated user interaction which provides “passive” simulated human interaction. The “passive” simulated human interaction is in accordance with the selected action profile, but it is not responsive to an input request by the launched object 147 (e.g., a behavior of the launched object that requiring user action). In some cases, the simulated human interaction is in response to a prescribed level of inactivity by the object.

Herein, the “passive” simulated human interaction may include any simulated operations that, without prompting, may be conducted by the user on the object such as moving to a particular page in a Microsoft® Office Word document (object) or switching to a particular tab in a Microsoft® Office Excel document (object). As an illustrative example, assuming the object has an object-type of a Microsoft® Office Excel document, experiential knowledge of typical placement of exploit/malicious code (e.g., through machine learning techniques) in a Microsoft® Office Excel document may result in instructions in the selected action profile for the passive UI simulation logic 370 to switch to the second sheet of the Microsoft® Office® Excel document at a predetermined time after the actuation logic launches the object.

The device control simulation logic 380 is a third type of simulated user interaction that may be performed during virtual analysis of the suspect object 147. The device control simulation logic 380 simulates device control interactions that are object-type agnostic. For example, the device control simulation logic 380 may receive instructions from the selected action profile 300₁ to simulate certain device control interactions, such as simulate particular keystrokes and/or particular mouse movements, in an attempt to detonate a malicious object that is awaiting user interaction before conducting a malicious attack.

Additionally, the UI framework log 176 records the activities conducted by the simulation logic 350. As discussed above, the UI framework log 176 may record any suspicious activity and/or malicious activity as well as any actions taken, or refrained from being taken, any requested input and timestamps for all actions and requested input. Upon completion of the dynamic analysis, the information recorded in the UI framework log 176 may be accessible to the score determination logic 168 and/or the classification engine 190.

It is contemplated that the action profile(s) 188 may be updated through a configuration file that may be propagated to the MDS 110₁ over a network 125 of FIG. 2. For example, the action profile update may be provided by the management system 220 over network 125 via machine learning engine 122, which receives information associated with malicious objects as reported by reporting engine 195 based on content within the UI framework log 176. Additionally, the machine learning engine 122 receives information associated with UI-dependent malicious objects as reported by MDSes (e.g., MDS 110₂ or MDS 110₃). The machine learning engine 122 utilizes this information, and information from third party sources, to develop action profile updates. The action profile update may include revised rules, new instructions or commands, and/or altered parameters that may provide improved malware detection by targeting characteristics in an object-type and/or targeting newly identified malware inclusive of exploits, suspicious code and/or malicious code or other data that assists in conducting a malicious attack on a network or network device.

Alternatively, the action profile update may be provided by over the network 105 (for example through a download using the cloud computing services 228 and/or manual installation through the use of a storage device such as flash storage).

IV. User Interaction Conducted by the Simulation Logic

Referring now to FIG. 4A, a flowchart illustrating an exemplary method conducted by the passive UI simulation logic and the device control simulation logic for detecting malware is shown. Each block illustrated in FIG. 4A represents an operation performed in accordance with a selected action profile for providing targeted, simulated user interaction during analysis of the object 147 within a virtual run-time environment. These operations are conducted by the MDS 100₁ in efforts to automatically, without human interaction, detect malicious objects that commence a malicious attack in response to human interaction. For this embodiment, with the selected active profile, “active” simulated human interaction responsive to input requests from the object is assigned the highest priority, while “passive” simulated human interactions and simulated device control interactions are assigned lesser priority.

Herein, a first determination is made as to whether the object has been launched by the actuation logic (block 400). If not, the UI framework does not receive a launch notification from the actuation logic, and thus, the simulation logic remains in an idle state. However, once an object is launched, the simulation logic receives a launch notification from the actuation logic, which causes the simulation logic to reference the selected action profile. A first determination is made as to whether user interaction is currently being requested based on resultant behaviors of the object during analysis (block 405). Stated differently, a determination is made as to whether the object process has initiated an input request, where timely “active” simulated human interaction is necessary. This determination may be conducted by monitoring system calls and other signaling that is directed to generation of a dialog box, text box, window or other perceivable element that would require user interaction.

In event that the passive UI simulation logic is currently conducting “passive” simulated human interactions and/or the device control simulation logic is currently conducting simulated device control interactions in accordance with rules outlined in the selected action profile, these simulated operations are paused for a prescribed duration. The prescribed duration may be set by the rules set forth in the selected action profile that identify the amount of time necessary to complete a particular type of “active” simulated human interaction. Furthermore, the “paused” simulated operations are time-stamped and placed in a wait queue for subsequent processing after the active UI simulation logic has completed its simulated human interaction. The selected action profile triggers the active UI simulation logic to conduct a particular “active” simulated human interaction and store the activity in the UI framework log. Thereafter, the simulation logic determines if the analysis of the object has completed, and if not, cycles back to determine whether the object is actively requesting user interaction (blocks 410-425)

In the event that active user interaction is not needed at this time, a determination is made as to whether there are any “paused” passive simulated human interactions and/or simulated device control interactions (blocks 405 and 430). This determination may be accomplished by analysis of the wait queue and/or determining whether a prescribed wait duration has elapsed (e.g., difference between current time and the time-stamp is greater than or equal to the prescribed duration). If so, these paused simulated operations are resumed (block 435). However, if there are no paused passive simulated human interactions and/or the simulated device control interactions, a determination is made as to whether there are any “passive” simulated human interactions that, according to the selected action profile, should be initiated (block 440). If so, the selected action profile triggers the passive UI simulation logic to conduct a particular “passive” simulated human interaction and store the activity in the UI framework log (block 445).

In the event that there has been at least a predetermined level of UI simulated activity thus far, the simulation logic may return to determine if the analysis of the object has completed, and if not, cycles back to determine whether the object is actively requesting certain user interaction (blocks 450, 420 and 405). The prescribed level of UI simulated activity may be measured by a variety of ways. For instance, the prescribed level of UI simulated activity may be determined based on whether simulated human interactions have occurred for a certain percentage of the run-time since the object was launched. Alternatively, the process may determine the number of “active” simulated human interactions or the number of active/passive simulated human interactions that have been completed since the object was launched.

In the event that the predetermined level of UI simulated activity has not been met, the device control simulation logic accesses the selected action profile to determine what simulated device control interactions are requested by the selection action profile, and thereafter, the selected action profile triggers the device control simulation logic to simulate such device controls and store such activity in the UI framework log. Thereafter, the simulation logic returns to determine if the analysis of the object has completed, and if not, cycles back to determine whether the object is actively requesting user interaction (blocks 450, 420 and 405).

According to these operations, the UI control logic is adapted to prioritize “active” simulated human interaction above “passive” simulated human interaction and the simulated device control interaction. Hence, in some cases as described herein, simulated human interaction and simulated device control interactions may be temporarily halted to direct resources to respond to an activity initiated by the object. Of course, it is contemplated that some types of “passive” simulated human interactions and simulated device control interactions may continue despite detection of an input request by the object. This may be done to maintain perceived consistency in simulated operations to avoid sophisticated malware to detect abnormally prompt changes in operation.

Referring now to FIG. 4B, a flowchart illustrating an exemplary method conducted by the active UI simulation logic for detecting malware is shown. As previously described, a first determination is made as to whether the object has been launched by the actuation logic (block 470). If not, the active UI simulation logic remains in an idle state. However, once an object is launched, the active UI simulation logic receives a notification from the actuation logic, where the simulation logic determines whether the suspect object is currently requesting some sort of active user interaction such as a dialog box is opened that required dismissal before the object continues its operation for example (block 475).

If so, the active UI simulation logic notifies the passive UI simulation and the device control simulation logic of an imminent active user interaction (block 480). This notification prompts the passive UI simulation and the device control simulation logic to pause any current operations as described in FIG. 4A. Thereafter, the active UI simulation logic performs a human simulation operation in accordance with rules set forth in the selected action profile and such activity is stored in the UI framework logic (block 485).

If the active UI simulation logic determines that the suspect object under analysis is not currently requesting active user interaction or responsive simulated human interactions have been provided, the active UI simulation logic determines whether the analysis of the suspect object has completed. If not, the active UI simulation logic initiates another iterative cycle awaiting a requested user interaction (block 490).

V. Software Deployment of the MDS

Referring now to FIG. 5, an exemplary block diagram of logic associated with the MDS 110₁ of FIG. 1 is shown. The MDS 110₁ comprises one or more processors 500 (hereinafter “processor(s)”), which is coupled to a first communication interface logic 510 via a first transmission medium 520. The first communication interface logic 510 may provide a communicative coupling with the network interface 136 of FIG. 1. Additionally, the processor(s) 500 may be communicatively coupled to a second communication interface logic 530 via a second transmission medium 540, which may provide communications with other MDSes 110₂-110₃ and management system 120 of FIG. 1.

According to one embodiment of the disclosure, the first communication interface logic 510 and/or the second communication interface logic 530 may be implemented as a physical interface including one or more ports for wired connectors. Additionally, or in the alternative, the first communication interface logic 510 and/or the second communication interface logic 530 may be implemented with one or more radio units for supporting wireless communications with other network devices.

The processor(s) 500 are further coupled to the persistent storage 550 via the transmission medium 560. According to one embodiment of the disclosure, the persistent storage 550 may be configured to store software components associated with the static analysis engine 145, the dynamic analysis engine 160, the classification engine 190 and the reporting engine 195. As shown, software components associated with the static analysis engine 145 may include the static analysis logic 152, the metadata extraction logic 154 and/or the object-type determination logic 156. The persistent storage 550 may be further configured to store software components associated with the dynamic analysis engine 160, which includes the VMM 172 along with the VMs 170₁-170_M. All or some of the VMs 170₁-170_M may be provisioned with the UI control logic 182, which may include the profile selector 184, UI framework 186 and/or action profile(s) 188.

Additionally, the persistent storage 550 may include the magic number database 158 that is accessed by the object-type determination logic 156 (described above) and data stores 159 and 164 that may operate, at least part, as data buffers.

In the foregoing description, the invention is described with reference to specific exemplary embodiments thereof. It will, however, 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.

Claims

1. A non-transitory computer readable storage medium having stored thereon logic that, upon execution by one or more processors implemented within a network device, performs operations comprising: launching, by an actuation logic, an object within a virtual run-time environment to detect a presence of malware or the object is part of a malicious attack; andselecting, by a control logic, one or more simulated user interactions with the object based on, at least in part, data within metadata associated with the object, the data identifying at least an object type corresponding to the object launched by the actuation logic.

2. The non-transitory computer readable storage medium of claim 1, wherein the launching of the object further comprises selecting the actuation logic based on the metadata, the actuation logic includes a software component implemented as part of a software profile that provisions one or more virtual machines operating within the virtual run-time environment.

3. The non-transitory computer readable storage medium of claim 2 further comprising: responsive to determining a first user interaction is being requested by the object, triggering a first simulated human interaction to occur, the first simulated human interaction being part of the one or more simulated user interactions;responsive to determining no user interaction is being requested by the object, resuming a second simulated human interaction, the second simulated human interaction having been previously triggered and paused upon determining the first user interaction is being requested.

4. The computer readable storage medium of claim 3, wherein the first simulated human interaction comprises a simulated action that is responsive to a behavior by the object having been launched by the actuation logic in the one or more virtual machines, the behavior represents an operating state of the object where the object is actively waiting for user input.

5. The computer readable storage medium of claim 4, wherein the first simulated human interaction comprises at least one of (i) a simulated action of closing a window or a dialog box, (ii) a simulated action of selecting a particular radio button, or (iii) a simulated action of entering one or more characters into a text box.

6. The computer readable storage medium of claim 4, wherein the second simulated human interaction comprises a simulated action that is initiated by a user during virtual analysis of the launched object.

7. The computer readable storage medium of claim 6, wherein the second simulated human interaction comprises at least one of (i) a simulated action of scrolling a page of a document corresponding to the object, (ii) a simulated action of selecting a certain tab of a spreadsheet corresponding to the object, (iii) a simulated action of selecting a particular page of the document, or (iv) a simulated action of accessing one or more menu options.

8. The computer readable storage medium of claim 4, wherein the second simulated human interaction comprises simulated device control for an input device that controls operations of an endpoint targeted to receive data including the object.

9. The computer readable storage medium of claim 8, wherein the simulated device control comprises at least one of (i) a simulated action of a keystroke for a keyboard corresponding to the input device, (ii) a simulated action of a movement of a mouse corresponding to the input device, (iii) a simulated action of a click of a button on the mouse, or (iv) a simulated action of an area of a touch screen corresponding to the input device.

10. The computer readable storage medium of claim 3, wherein the first simulated human interaction is controlled by a first user interaction (UI) simulation logic and the second simulated human interaction is controlled by a second UI simulation logic, the second UI simulation logic is aware of the first simulated human interaction being performed by the first UI simulation logic and pauses the second simulated human interaction in favor of the first simulated human interaction for later resumption of the second simulated human interaction.

11. The computer readable storage medium of claim 1 further comprising: prior to launching the object within the virtual run-time environment, selecting an action profile by a profile selector based on the received metadata associated with the object, the action profile comprises a plurality of rules that dynamically control the one or more simulated user interactions.

12. The computer readable storage medium of claim 11, wherein the action profile is selected by either (i) the profile selector provisioned within one or more virtual machines operating within the virtual run-time environment or (ii) the profile selector implemented within a virtual machine monitor (VMM) communicatively coupled to one or more virtual machines operating within the virtual run-time environment.

13. The computer readable storage medium of claim 1, wherein the selecting of the one or more simulated user interactions with the object is further based on, at least in part, at least one of (i) data identifying whether the object is encrypted and a type of encryption, (ii) data identifying whether the object is or contains an embedded object, or (iii) data identifying whether the object includes password-protected fields.

14. An apparatus for detecting malware with an object, the apparatus comprising: one or more processors; anda storage device communicatively coupled to the one or more processors, the storage device comprises one or more action profiles, each action profile of the one or more action profiles being a collection of instructions or commands that performs user interaction (UI) activity in accordance with a set of rules prescribed for the corresponding action profile,a profile selector for selecting an action profile from the one or more action profiles based on metadata associated with the object, the metadata includes data identifying a type of the object, anda UI framework logic that, in response to the object being launched within a virtual machine, performs simulated user interactions with the object in accordance with the set of rules prescribed in the selected action profile, the UI framework logic including (i) an actuation logic to launch the object, and (ii) simulation logic to dynamically control the simulated user interactions conducted on the launched object.

15. The apparatus of claim 14, wherein the simulation logic comprises (i) an active UI simulation logic, (ii) a passive UI simulation logic, and (iii) a device control simulation logic.

16. The apparatus of claim 15, wherein the active UI simulation logic is configured to: detect an input request initiated by the launched object, andresponsive to detecting the input request initiated by the launched object, determine a response based on contents of the selected action profile.

17. The apparatus of claim 16, wherein the input request includes an attempted display of a dialog box.

18. The apparatus of claim 17, wherein the response comprises at least one of (i) a simulated action of closing the dialog box, and (ii) a simulated action of selecting a particular radio button associated with the dialog box.

19. The apparatus of claim 15, wherein the passive UI simulation logic is configured to provide a simulated human interaction during virtual analysis of the launched object, the simulated human interaction represents user-initiated simulated actions.

20. The apparatus of claim 15, wherein the device control simulation logic is configured to simulate device control interactions that are object-type agnostic.

21. The apparatus of claim 14, wherein the storage device further comprising a log including data so that, when the launched object is classified as malware, the data is used to update the one or more action profiles by indicating a set of simulated user interactions that lead to successful triggering of malicious behavior by the launched object.

22. The apparatus of claim 14, wherein the storage device further comprising logic to transmit the data within the log to a cloud infrastructure when the launched object is classified as suspicious as a result of virtual analysis, wherein the one or more action profiles are updated based on a lack of detonation of malware within the launched object or an obstruction of the detonation of malware within the launched object.

23. A computerized method implemented with a network device for detecting malware, comprising: launching, by an actuation logic within the network device, an object within a virtual run-time environment to detect a presence of malware or the object is part of a malicious attack; andselecting, by a profile selector, an action profile based on metadata associated with the object, wherein the metadata comprises data identifying an object type corresponding to the object and the action profile comprises a set of rules that dynamic control one or more simulated user interactions with the launched object.

24. The method of claim 23, wherein the launching of the object further comprises selecting the actuation logic based on the metadata.

25. The method of claim 23, wherein the simulated user interactions comprising: responsive to determining a first user interaction is being requested by the launched object, triggering a first simulated human interaction to occur, the first simulated human interaction being part of the one or more simulated user interactions;responsive to determining no user interaction is being requested by the object, resuming a second simulated human interaction, the second simulated human interaction having been previously triggered and paused upon determining the first user interaction is being requested.

26. The method of claim 25, wherein the first simulated human interaction comprises a simulated action that is responsive to a behavior by the launched object in one or more virtual machines that are part of the run-time virtual environment, the behavior represents an operating state of the launched object where the launched object is actively waiting for user input.

27. The method of claim 25, wherein the second simulated human interaction comprises a simulated action that is initiated by a user during virtual analysis of the launched object.

28. The method of claim 25, wherein the second simulated human interaction comprises simulated device control for an input device that controls operations of an endpoint targeted to receive data including the object.

29. The method of claim 25, wherein the first simulated human interaction is controlled by a first user interaction (UI) simulation logic and the second simulated human interaction is controlled by a second UI simulation logic, the second UI simulation logic is aware of the first simulated human interaction being performed by the first UI simulation logic and pauses the second simulated human interaction in favor of the first simulated human interaction for later resumption of the second simulated human interaction.

30. The method of claim 23, wherein the action profile is selected from a plurality of action profiles by the profile selector provisioned within one or more virtual machines operating within the virtual run-time environment.

31. The method of claim 30, wherein the plurality of action profiles are updated via network delivered updates to modify instructions, parameters or rules within an action profile to be updated.

32. The method of claim 23, wherein the action profile is selected from a plurality of action profiles by the profile selector implemented within a virtual machine monitor (VMM) communicatively coupled to one or more virtual machines operating within the virtual run-time environment.

33. The method of claim 25 further comprising: recording, in a log, selection of the action profile, determination of the first user interaction requested by the object, triggering of the first simulated human interaction, and triggering of the second simulated human interaction.

34. The method of claim 33, wherein data stored within the log is subsequently used to update a plurality of action profiles including the selected action profile by indicating whether certain simulated user interactions detonated malware within the object.