System and method for detecting time-bomb malware

Abstract

According to one embodiment, a system comprises one or more counters; comparison logic; and one or more hardware processors communicatively coupled to the one or more counters and the comparison logic. The one or more hardware processors are configured to instantiate one or more virtual machines that are adapted to analyze received content, where the one or more virtual machines are configured to monitor a delay caused by one or more events conducted during processing of the content and identify the content as including malware if the delay exceed a first time period.

Description

FIELD

Embodiments of the disclosure relate to the field of data security. More specifically, one embodiment of the disclosure relates to a system, apparatus and method that enhances detection of time-bomb malware, namely malware with delayed activation.

GENERAL BACKGROUND

Over the last decade, malicious software (malware) attacks have become a pervasive problem for Internet users and enterprise network administrators. In most situations, malware is a program or file that is embedded within downloadable content and designed to adversely influence or attack normal operations of an electronic device (e.g. computer, tablet, smartphone, server, router, wearable technology, or other types of products with data processing capability). Examples of different types of malware may include bots, computer viruses, worms, Trojan horses, spyware, adware, or any other programming that operates within an electronic device without permission by the user or a system administrator.

Over the past few years, various types of security appliances have been deployed within an enterprise network in order to detect behaviors that signal the presence of malware. Some of these security appliances conduct dynamic analysis on suspicious content within a sandbox environment in order to determine if malware is present. As a result, some malware is now being coded to evade analysis within a sandbox environment.

Currently, there are various techniques that malware is using to evade sandboxed malware analysis. They can be broadly categorized as:

• • [1] Environment checking: Malware checks for several environmental facts to identify whether it is being run in a sandbox. In response, the malware may halt its execution to avoid detection upon sandbox detection. This may be accomplished by the malware querying for a CPUID string;
  • [2] User Interaction: Malware will not perform any malicious activity until some user interaction is provided, contrary to capabilities of most sandbox environments.
  • [3] Presence of AV/Detection tool: Malware checks for specific artifacts that indicate an anti-virus or sandboxed detection is in effect (e.g. if certain system APIs are hooked); and
  • [4] Stalling: Malware delays execution for substantial time such that the malicious activity is not performed within run-time of sandbox.

As a result, mechanisms are necessary to detect all types of malware, even malware that is specifically configured to evade detection within a sandbox environment such as a virtual machine (VM) based environment.

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. 1 is a first exemplary block diagram of a malware content detection (MCD) system with for multiple VMs deployed within a communication network.

FIG. 2 is a detailed exemplary embodiment of the MCD system of FIG. 1.

FIG. 3 is a second exemplary block diagram of the MCD system of FIG. 1.

FIG. 4 is a detailed exemplary block diagram of components within the hardware architecture of the MCD system of FIG. 1.

FIG. 5 is an exemplary embodiment of a flowchart illustrating operations for monitoring sleep operations conducted within the VM-based analysis environment of FIG. 2.

FIG. 6 is an exemplary embodiment of a flowchart illustrating operations in monitoring call operations conducted within the VM-based analysis environment of FIG. 2.

FIG. 7 is an exemplary embodiment of a flowchart illustrating operations in monitoring an instruction profile associated with operation conducted within the VM-based analysis environment of FIG. 2.

DETAILED DESCRIPTION

Various embodiments of the disclosure relate to a system, apparatus and method for enhancing detection of malware that is configured to avoid detection within a sandbox environment, such as a virtual machine based (VM-based) analysis environment for example, through delayed activation. Such malware is commonly referred to as a “time-bomb malware”. This enhanced detection may be conducted by tracking call site time delays and access frequency, which may be captured inside a VM in user-mode or kernel mode, inside of the virtual (execution) environment, or even external to the VM environment (e.g., frequency based threshold detection logic) and utilized to evade automated analysis environments.

One embodiment of the disclosure is directed to logic that is configured to monitor one or more operations within a VM-based analysis environment and, where applicable, adjusts one or more parameters associated with such operations. These operating parameters may involve requests and/or calls that delay further processing of content loaded into the VM-based analysis environment. For example, the operating parameter may include, but are not limited or restricted to (i) the number of Sleep request messages; (ii) an amount of time requested for a particular Sleep request message; (iii) the cumulative amount of Sleep time requested over a predetermined run time; (iv) the number of function calls from a particular call site to an API (e.g., addressed function such as a specific Application Programming Interface “API”); (v) the total frequency of the instruction pointer remaining within a particular address range; (vi) processor utilization level; or the like.

Hence, according to one embodiment of the disclosure, one or more counters may be deployed for monitoring the number of Sleep request messages initiated by the content under analysis. The content is determined to be associated with time-bomb malware if the number of Sleep request messages initiated by content under analysis exceeds a first threshold value (e.g. predetermined time value). Similarly, one or more counters may be deployed for monitoring whether the cumulative Sleep time for multiple Sleep request messages initiated by the content under analysis exceeds a second threshold value. If so, the content is determined to be associated with time-bomb malware.

Also, in the alternative or in combination with the sleep counter(s), one or more counters may be deployed for monitoring the total number of function calls initiated by the content under analysis. The content is determined to be associated with time-bomb malware if the total number of function calls exceeds a third threshold value. Similarly, one or more counters may be deployed for monitoring the number of function calls directed from a particular call site to an API by the content under analysis, where the content is determined to be associated with time-bomb malware if the number of function calls to the particular API exceed a fourth threshold.

According to a further embodiment of the disclosure, logic may be deployed to work in concert with each VM to monitor whether the content under analysis is repeatedly executing instructions located at a specific address or address range, which denotes a programming “loop” operation. If so, the content is determined to be associated with time-bomb malware.

In summary, multiple APIs and certain assembly instructions may be utilized by time-bomb malware to get OS time information. Also, the API access patterns for time delay purposes may vary from one malicious sample to another. For instance, the time-bomb malware may be a tightly loop execution, repeatedly performing GetLocalTime, Compare and/or Sleep calls, where detonation of the time-bomb malware occurs upon reaching a desired date. Other time-bomb malware may be a mixture of Sleep calls, Floating-Point Math calls, and/or subroutines that also stall and sleep. Yet another time-bomb malware may involve a tight execution loop of GetLocalTime, Compare, Sleep, and/or Cut-and-Paste operations, which identifies the importance of identifying the accessed call-site of the delay-execution code, so that proper threshold based statistics logic may be applied to multiple variants.

Herein, the disclosure describes different embodiments for addressing certain types of stall technique such as Sleep calls or the like. It is contemplated that the scope of the invention is directed to a mechanism that detects time-bomb malware associated with not only repeated Sleep calls but with any event or call variant adapted to stall malware execution to evade a sandbox environment.

I. Terminology

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

Logic (or engine) also may be 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.

The term “content” generally refers to information, such as text, software, images, audio, metadata and/or other digital data for example, that is transmitted as one or more messages. Each message(s) may be in the form of a packet, a frame, an Asynchronous Transfer Mode “ATM” cell, or any other series of bits having a prescribed format. The content may be received as a data flow, namely a group of related messages, being part of ingress data traffic.

One illustrative example of content includes web content, namely data traffic that may be transmitted using a Hypertext Transfer Protocol (HTTP), Hypertext Markup Language (HTML) protocol, or any other manner suitable for display on a Web browser software application. Another example of content includes one or more electronic mail (email) messages, which may be transmitted using an email protocol such as Simple Mail Transfer Protocol (SMTP), Post Office Protocol version 3 (POP3), or Internet Message Access Protocol (IMAP4). Yet another example of content includes an Instant Message, which may be transmitted using Session Initiation Protocol (SIP) or Extensible Messaging and Presence Protocol (XMPP) for example. A final example of content includes one or more files that are transferred using a data transfer protocol such as File Transfer Protocol (FTP) for subsequent storage on a file share.

The term “time-bomb malware” is software that includes at least one exploit, namely a particular portion of software that, after intentional delayed execution, takes advantage of one or more vulnerabilities within system software and produces an undesired behavior. The behavior is deemed to be “undesired” based on customer-specific rules, manufacturer-based rules, or any other type of rules formulated by public opinion or a particular governmental or commercial entity. Examples of an undesired behavior may include a communication-based anomaly or an execution-based anomaly that (i) alters the functionality of an electronic device and/or (ii) provides an unwanted functionality which may be generally acceptable in other context.

The term “transmission medium” is a communication path between two or more systems (e.g. any electronic devices with data processing functionality such as, for example, a security appliance, server, mainframe, computer, netbook, tablet, smart phone, router, switch, bridge or brouter). The communication path may include wired and/or wireless segments. Examples of wired and/or wireless segments include electrical wiring, optical fiber, cable, bus trace, or a wireless channel using infrared, radio frequency (RF), or any other wired/wireless signaling mechanism.

In general, a “virtual machine” (VM) is a simulation of an electronic device (abstract or real) that is usually different from the electronic device conducting the simulation. VMs may be based on specifications of a hypothetical electronic device or emulate the architecture and functions of a real world computer. A VM can be one of many different types such as, for example, hardware emulation, full virtualization, para-virtualization, and/or operating system-level virtualization virtual machines.

A “software profile” is information that is used for virtualization of an operating environment (e.g. instantiation of a VM) that is adapted to receive content for malware analysis. The software profile may identify a guest operating system “OS” type; a particular version of the guest OS; one or more different application types; particular version(s) of the application type(s); virtual device(s); or the like.

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.

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. Time-Bomb Malware Detection Architecture

Referring to FIG. 1, an exemplary block diagram of a communication system 100 deploying a plurality of malware content detection (MCD) systems 110₁-110_N (N>1, e.g. N=3) communicatively coupled to a management system 120 via a network 125 is shown. In general, management system 120 is adapted to manage MCD systems 110₁-110_N. For instance, management system 120 may be adapted to cause malware signatures generated as a result of time-bomb malware detection by any of MCD systems 110₁-110_N (e.g. MCD system 110₂) to be shared with one or more of the other MCD systems 110₁-110_N (e.g. MCD system 110₁) including where such sharing is conducted on a subscription basis.

Herein, according to this embodiment of the disclosure, first MCD system 110₁ is an electronic device that is adapted to (i) intercept data traffic routed over a communication network 130 between at least one server device 140 and at least one client device 150 and (ii) monitor, in real-time, content within the data traffic. More specifically, first MCD system 110₁ may be configured to inspect content received via communication network 130 and identify “suspicious” content. The incoming content is identified as “suspicious” when it is assessed, with a certain level of likelihood, that at least one characteristic identified during inspection of the content indicates the presence of an exploit.

Herein, according to one embodiment of the disclosure, the first MCD system 110₁ is a web-based security appliance configured to inspect ingress data traffic and identify whether content associated with the data traffic includes time-bomb malware. The communication network 130 may include a public computer network such as the Internet, in which case an optional firewall 155 (represented by dashed lines) may be interposed between communication network 130 and client device(s) 150. Alternatively, the communication network 130 may be a private computer network such as a wireless telecommunication network, wide area network, or local area network, or a combination of networks.

The first MCD system 110₁ is shown as being coupled with the communication network 130 (behind the firewall 155) via a network interface 160. The network interface 160 operates as a data capturing device (referred to as a “network tap”) that is configured to receive data traffic propagating to/from the client device(s) 150 and provide some or all of the content associated with the data traffic (e.g. objects) to the first MCD system 110₁.

In general, the network interface 160 receives and copies the content that is received from and provided to client device 150. Alternatively, the network interface 160 may copy only a portion of the content, for example, a particular number of objects associated with the content. For instance, in some embodiments, the network interface 160 may capture metadata from data traffic intended for client device 150, where the metadata is used to determine (i) whether content within the data traffic includes any exploits and/or (ii) the software profile used instantiating the VM(s) for exploit detection on the content.

It is contemplated that, for any embodiments where the first MCD system 110₁ is implemented as an dedicated appliance or a dedicated computer system, the network interface 160 may include an assembly integrated into the appliance or computer system that includes network ports, network interface card and related logic (not shown) for connecting to the communication network 130 to non-disruptively “tap” data traffic propagating therethrough and provide a copy of the data traffic to the heuristic engine 170 of MCD system 110₁. In other embodiments, the network interface 160 can be integrated into an intermediary device in the communication path (e.g. firewall 155, router, switch or other network device) or can be a standalone component, such as an appropriate commercially available network tap. In virtual environments, a virtual tap (vTAP) can be used to copy traffic from virtual networks.

Referring to both FIGS. 1 and 2, first MCD system 110₁ comprises a heuristic engine 170, a heuristics database 175, an analysis engine 180, a scheduler 185, a storage device 190, and a reporting module 195. In some embodiments, the network interface 160 may be contained within the first MCD system 110₁. Also, heuristic engine 170, analysis engine 180 and/or scheduler 185 may be hardware logic implemented with a processor or other types of circuitry. Alternatively, this logic may be configured as software modules executed by the same or different processors. As an example, the heuristic engine 170 may be one or more software modules executed by a first hardware processor implemented within the first MCD system 110₁, while the analysis engine 180 and/or scheduler 185 may be executed by a second hardware processor. These processors may be located at geographically remote locations and communicatively coupled via a network.

In general, the heuristic engine 170 serves as a filter to permit subsequent malware analysis on portion(s) of incoming content that may have time-bomb malware. As an ancillary benefit, by analyzing only the portion of the incoming content that may have such malware, various system resources may be conserved and a faster response time may be provided in determining the presence of malware within analyzed content.

As still shown in FIG. 1, the heuristic engine 170 receives the incoming content from the network interface 160 and applies heuristics to determine if any of the content is “suspicious”. The heuristics applied by the heuristic engine 170 may be based on data and/or rules stored in the heuristics database 175. Also, the heuristic engine 170 may examine the image of the captured content without executing or opening the captured content.

For example, the heuristic engine 170 may examine the metadata or attributes of the captured content and/or the code image (e.g., a binary image of an executable) to determine whether the captured content matches or has a high correlation with a predetermined pattern of attributes that is associated with a malicious attack, especially time-bomb malware attacks. According to one embodiment of the disclosure, the heuristic engine 170 flags content from one or more data flows as suspicious after applying this heuristic analysis.

It is contemplated that the heuristic engine 170 may comprise a static analysis tool 175 that is configured to parse malware binaries and specifically identify delay hotspots. A “delay hotspot” could be an API call or Sleep request with large timeout values or it could be a loop with high repeat counter. Such analysis can complement dynamic analysis technique and lead to more definitive detection of time-bomb malwares.

Thereafter, according to one embodiment of the disclosure, the heuristic engine 170 may be adapted to transmit at least a portion of the metadata or attributes of the suspicious content 172, which may identify attributes of the client device 150, to a control logic 182 implemented within analysis engine 180. Such metadata or attributes are used to identify software profile information used to instantiate at least one VM for subsequent malware analysis. In another embodiment of the disclosure, the control logic 182 may be adapted to receive one or more messages (e.g. data packets) from the heuristic engine 170 and analyze the message(s) to identify the software profile information for instantiating the VM(s) subsequently used for malware analysis.

Control logic 182 is adapted to control formation of one or more VM-based analysis environments 200₁-200_M as shown in FIG. 2. As shown herein, at least one analysis environment 200₁ comprises a VM 210₁ with corresponding sleep analysis logic 220 (e.g., one or more sleep counters, etc.) and call analysis logic 230 (e.g., one or more call counters, etc.), comparison logic 240 (e.g. one or more comparators, etc.) along with instruction pointer analysis logic (profiler) 250 and processor (CPU) statistic monitoring logic 260.

For instance, as an illustrative example, the suspicious content under analysis may include an email message that was generated, under control of Windows® 7 Operating System, using a Windows® Outlook 2007, version 12. The email message further includes a Portable Document Format (PDF) attachment in accordance with Adobe® Acrobat®, version 9.0. Upon determining that the email message includes suspicious content, heuristic engine 170 and/or control logic 182 may be adapted to provide software profile information to scheduler 185 in order to identify a particular type of VM needed to conduct dynamic analysis of the suspicious content. According to this illustrative example, the software profile information would identify the VM software as (1) Windows® 7 Operating System (OS); (2) Windows® Outlook 2007, version 12; and (3) Adobe® Acrobat® PDF reader that allows viewing of the above-identified PDF document.

The control logic 182 supplies the software profile information to the scheduler 185, which conducts a search of information within storage device 190 to determine if a VM image 192 identified by the software profile information resides within storage device 190. The VM image 192 supports the appropriate OS (e.g. Windows® 7 OS) and one or more applications (e.g., Windows® Outlook 2007, version 12; and Adobe® Acrobat® PDF reader). If so, the scheduler 185 uses that the VM image 192 to instantiate a VM within analysis environment 200₁ in order to analyze the suspicious content and determine if such content is associated with time-bomb malware.

Of course, it is contemplated that if the storage device 190 does not feature a software profile supporting the above-identified OS/application(s), the scheduler 185 may simply ignore the VM request from control logic 182 or may obtain an VM image directed to similar software. For example, the scheduler 185 may be adapted to obtain a VM image based on the same OS but a different version(s) of the targeted application(s). Alternatively, the scheduler 185 may be adapted to obtain the same OS (e.g. Windows® OS 7) along with an application different from the targeted application but having similar functionality. As another alternative, the scheduler 185 may receive a different OS image that supports similar functionality.

In another embodiment of the disclosure, the heuristic engine 170 may determine the software profile information from the data traffic by receiving and analyzing the content from the network interface 160. For instance, according to one embodiment of the disclosure, it is contemplated that the heuristic engine 170 may be adapted to transmit the metadata identifying the client device 150 to the analysis engine 180, where such metadata is used to identify a desired software profile. The heuristic engine 170 may then transmit the software profile information to a scheduler 185 in lieu of such information being provided from control logic 182 within the analysis engine 180.

Alternatively, the control logic 182 may be adapted to receive one or more data packets of a data flow from the heuristic engine 170 and analyze the one or more data packets to identify the software profile without pre-processing by heuristic engine 170. In yet other embodiment of the disclosure, the scheduler 185 may be adapted to receive software profile information, in the form of metadata or data packets, from the network interface 160 or from the heuristic engine 170 directly.

The storage device 190 may be configured to store one or more VM disk files forming a VM profile database 194, where each VM disk file is directed to a different software profile for a VM. In one example, the VM profile database 194 may store a plurality of VM disk files having VM images for multiple software profiles in order to provide the collective capability for simulating the performance of a wide variety of client device(s) 150.

The analysis engine 180 is adapted to execute multiple VMs concurrently to support different VM operating environments that simulate the receipt and/or processing of different data flows of “suspicious” content by different network devices. Furthermore, the analysis engine 180 analyzes the effects of such content during processing. The analysis engine 180 may identify exploits by detecting undesired behavior caused by simulated processing of the suspicious content as carried out by the VM. This undesired behavior may include numerous repeated functions calls, repeated Sleep calls, and other behavior to stall or delay execution of code associated with the incoming content.

The analysis engine 180 may flag the suspicious content as malware according to observed undesired behavior of the VM. Different types of behaviors may be weighted based on the likelihood of system compromise, where suspicious content is determined when the weighted value exceeds a certain threshold.

Of course, it is contemplated that, for deeper analysis to detect exploits, such operations may be conducted within the cloud 165 in lieu of or in addition to operations performed within analysis engine 180.

The reporting module 195 may issue alert messages indicating the presence of one or more exploits to one or more hardware processors executing outside the VM environments, and may use pointers and other reference information to identify what message(s) (e.g. packet(s)) of the suspicious content may contain the exploit(s). Additionally, the server device(s) 140 may be added to a list of malicious network content providers, and future network transmissions originating from the server device(s) 140 may be blocked from reaching their intended destinations, e.g., by firewall 155.

Referring now to FIG. 3, a second exemplary embodiment of MCD system 110₁ set forth in FIG. 1 is shown, where the software profile for VM instantiation is not determined through analysis of suspicious content (e.g. metadata, data packets, binary, etc.) by the network interface 160, heuristic engine 170, or analysis engine 180. Rather, this software profile directed to software under test is uploaded by the user and/or network administrator.

More specifically, a user interface 310 allows the user or network administrator (hereinafter referred to as “user/administrator”) to introduce objects 300 of the suspicious content in accordance with one or more prescribed software profiles 320. The prescribed software profile(s) 320 may be preloaded or selected by the user/administrator in order to instantiate one or more VMs based on operations of the scheduler 185 and storage device 190 as described above. The VMs perform dynamic analysis of the objects 300 to monitor for undesired behavior during virtual processing of these objects 300 within the VMs.

Referring now to FIG. 4, an exemplary block diagram of logic that is implemented within MCD system 110₁ is shown. MCD system 110₁ comprises one or more processors 400 that are coupled to communication interface logic 410 via a first transmission medium 420. Communication interface logic 410 enables communications with MCD systems 110₂-110_N of FIG. 1 as well as other electronic devices over private and/or public networks. According to one embodiment of the disclosure, communication interface logic 410 may be implemented as a physical interface including one or more ports for wired connectors. Additionally, or in the alternative, communication interface logic 410 may be implemented with one or more radio units for supporting wireless communications with other electronic devices.

Processor 400 is further coupled to persistent storage 430 via transmission medium 425. According to one embodiment of the disclosure, persistent storage 430 may include stalled processing analysis logic 440 and a data store 475. In general, stalled processing analysis logic 440 is configured to monitor and alter operating parameters for one or more VM-based analysis environments in order to improve reliability in detecting time-bomb malware. The results of the analysis are stored within data store 475.

More specifically, stalled processing analysis logic 440 comprises sleep analysis logic 220, call analysis logic 230, comparison logic 240, time adjustment logic 450, instruction pointer analysis logic 250, and processor statistic monitoring logic 260. Herein, both sleep analysis logic 220 and call analysis logic 230 are configured to address an event where a series of successive Sleep function calls are used to stall processing of the incoming content. As an example, sleep analysis logic 220 may be configured to monitor the number of Sleep calls, the Sleep intervals and the cumulative Sleep time. The call analysis logic 230 may be configured to perform the same general functionality in monitoring the number of function calls made globally or to a particular call site as well as the cumulative delay incurred by initiation of function calls.

Optionally working in concert with sleep analysis logic 220 and call analysis logic 230, the time adjustment logic 450 is configured to alter the time interval returned back to the content running in the VM-based analysis environment (e.g., environment 200₁). This alteration is performed in order to accelerate the time-bomb malware activity such that “suspicious” behavior can be detected. This is achieved through a shortened time interval while executing successive Sleep calls and/or other types of calls used when processing the content.

More specifically, sleep analysis logic 220 is configured with one or more counters that are used to count the number of Sleep request messages initiated by the content under analysis for subsequent comparison, using comparison logic 240 (e.g. one or more comparators) of the count value with a first threshold value stored in data store 475.

Additionally, the sleep analysis logic 220 may be further configured with one or more counters that are used to compute the cumulative amount of time (e.g. in units of time, clock cycles, etc.) in which the content under analysis would have placed a targeted electronic device into a Sleep state. The cumulative amount of time is subsequently compared, using comparison logic 240, to a second threshold value that is different than the first threshold value. The second threshold value may be set to a time value less than the average amount of time permitted for analysis of the content within a VM-based analysis environment 210₁. The content under analysis is considered to be associated with time-bomb malware if the first and/or second thresholds are exceeded.

Call analysis logic 230 is configured with one or more counters that are used to count the number of function calls initiated by the content under analysis, which is subsequently compared with a third threshold value stored in data store 475 using comparison logic 240. The number of function calls may be based on either (i) a global basis (e.g., total number of function calls) or (ii) a per call site basis (e.g. per each caller address). It is contemplated that the third threshold value may differ from the first and second threshold values, and the value may be based at least in part on the type of function call. For instance, the GetLocalTime function call may be analyzed with greater scrutiny as this API function call tends to be a common choice for repetitive call type of evasion. Other such APIs may include SystemTimeToFileTime, GetSystemTime, Process32First, NtYieldExecution, NtDelayExecution, SleepEx, and different Wait variants (e.g., MsgWaitForMultipleObjects, WaitForSingleObject, etc.).

Call analysis logic 230 may be further configured with one or more counters that are used to monitor the cumulative amount of time that the called functions would need for execution of the called function. Using comparison logic 240, the cumulative amount of time is subsequently compared to a fourth threshold value stored in data store 475. The content under analysis is associated with time-bomb malware if the third and/or fourth thresholds are exceeded.

Call analysis logic 230 also has the ability to report the calling module name (EXE/DLL) based on call site addresses. This allows the reporting module (195) to assign weights intelligently (e.g. less weight when a call site is from system module and higher weight when a call site is from the content under analysis).

As further shown in FIG. 4, time adjustment logic 450 is configured to operate in concert with sleep analysis logic 220 and/or call analysis logic 230 in order to compute a shortened time to be used in Sleep so that malware is forced to perform malicious activity within the VM analysis time duration.

The instruction pointer analysis logic 250 is configured to periodically check, during processing of the content under analysis, whether the instruction pointer has remained within one or more prescribed address range(s) over a prolonged period of time. This check is conducted in order to determine if the content includes time-bomb malware adapted to perform an instruction-based loop to evade analysis within the VM-based environment. If the instruction pointer analysis logic 250 determines that the instruction pointer continues to remain within a particular address range, the processor utilization measured by processor statistic monitoring logic 260 is greater than a prescribed value, and no other exploits have been detected, the instruction pointer analysis logic 250 determines that the content under analysis is associated with time-bomb malware.

According to one embodiment of the disclosure, at least the sleep analysis logic 220, call analysis logic 230 and time adjustment logic 450 are implemented as part of the VM. The comparison logic 240, instruction pointer analysis logic 250 and processor statistic monitoring logic 260 may be placed within the VM or outside the VM.

As additional counter measures to time-bomb malware, persistent storage 430 may include dynamic threshold generation logic 480 and/or call site management logic 490. Dynamic threshold generation logic 480 comprises logic that dynamically alters the threshold values utilized by the sleep analysis logic 220 and call analysis logic 230. The dynamic nature of the threshold values prevents malware writers from altering malware to circumvent established thresholds, if such thresholds are discovered.

Another optional logic implemented within MCD system 110₁ is the call site management logic 490. The call site management logic 490 is configured to maintain a finite number of call sites as part of a table. If the table is full and a request for a new call site is made, the call site management logic 490 determines if the new call site is associated with a larger processing time requirement than another call site within the table. If so, the new call site is substituted for that call site. If not, the new call site is not placed within the table. However, cumulative threshold is updated accordingly.

III. Time-Bomb Malware Detection Operations

Referring to FIG. 5, a first exemplary flowchart outlining the operations for time-bomb malware detection is shown. Upon receiving content, a determination is made as to whether the content is “suspicious,” namely whether analysis of the content indicates the presence of an exploit (blocks 500 and 510). Where the content is determined to be “suspicious,” the attributes of the content may be used to determine one or more software profiles (block 520). VMs within the analysis environment are based on these software profile(s).

Thereafter, the VM(s) perform operations on the suspicious content and analyzes the results of these operations to determine if any exploits are present (block 530). These operations may include Sleep analysis, Call analysis and profiling (e.g. processor utilization, addressing analysis, etc.) as described above. If no exploits are detected, no further time-bomb analysis is needed (block 540).

Otherwise, according to one embodiment of the disclosure, one or more counters are initiated during run-time of the content under analysis. The counter(s) may monitor (i) the number of Sleep request messages, (ii) the Sleep interval requested and/or (iii) the cumulative Sleep time (herein generally referred to as “Sleep activity”). Where the Sleep activity exceeds a prescribe threshold, a determination is made that the content under analysis includes time-bomb malware (blocks 550 and 560). As a result, the sleep analysis logic is adapted to emulate compliance with requested Sleep calls, where actual duration of the request Sleep time(s) is shortened by the time adjustment logic, in some cases significantly shortened 80%, 90% or more for the allocated sleep time (block 570). Such shortening of the Sleep time, which is unbeknownst to the content under analysis, alters the processing time frame for the VM environment and allows the VM to monitor and report the particulars behind the time-bomb malware attack. If the Sleep activity remains below the prescribed threshold, the VM continues to operate as normal (block 580).

Referring now to FIG. 6, a second exemplary flowchart outlining the operations for time-bomb malware detection is shown. Similarly, upon receiving content, a determination is made as to whether the content is “suspicious” based on a potential presence of an exploit (blocks 600 and 610). Where the content is determined to be “suspicious,” the attributes of the content may be used to determine one or more software profiles, where the VMs within the analysis environment are based on these software profile(s) (block 620).

Thereafter, the VM(s) perform operations on the suspicious content and analyzes the results of these operations to determine if any exploits are present (block 630). If no exploits are detected, no further time-bomb analysis is needed (block 640).

Otherwise, according to one embodiment of the disclosure, one or more counters are initiated during run-time of the content under analysis. The counter(s) may monitor the number of repeated function calls to a particular API. Where the number of function calls exceeds a prescribe threshold, a determination is made that the content under analysis includes time-bomb malware (blocks 650 and 660). As a result, the call analysis logic is adapted to emulate compliance with requested function calls by responding to these function calls, sometimes with a shortened call response wait time (block 670). Such shortened response time, which is unbeknownst to the content under analysis, alters the processing time frame for the VM environment to allow the VM to monitor and report the particulars behind the time-bomb malware attack. If the number of function calls to a particular API does not exceed a prescribed threshold, the VM will continue to operate as normal (block 680).

Referring to FIG. 7, a third exemplary flowchart outlining the operations for time-bomb malware detection is shown. Upon receiving content, a determination is made as to whether the content is “suspicious” based on a potential presence of an exploit (blocks 700 and 710). Where the content is determined to be “suspicious,” the attributes of the content may be used to determine one or more software profiles, where the VMs within the analysis environment are based on these software profile(s) (block 720).

Thereafter, the VM(s) perform operations on the suspicious content and analyzes the results of these operations to determine if any exploits are present (block 730). If no exploits are detected, no further time-bomb analysis is needed (block 740).

Otherwise, according to one embodiment of the disclosure, the instruction pointer analysis logic (profiler) undergoes operations to determine if the instruction pointer utilized during processing of the content under analysis is frequently located into the same memory address or a particular range of memory addresses (block 750). If not, no time-bomb malware is detected by the VM operating within the analysis environment of the analysis engine.

In the event that the profiler detects continued presence of the instruction pointer as described above, a determination is made by the processor statistics monitoring logic profile if processor utilization is greater than a prescribed operating threshold (blocks 760-770). If so, and no other malicious activity is detected, a determination is made that the content under analysis includes time-bomb malware (block 780). If the processor utilization is below the prescribed threshold or other malicious activity is seen, then no time-bomb malware is detected. Hence, the VM will continue to operate as normal (block 790).

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 system adapted with one or more processors and a non-transitory storage medium communicatively coupled to the one or more processors that are configured to instantiate a virtual machine that is adapted to receive content and process the received content, the system comprising: analysis logic configured to monitor one or more events representing operations within the virtual machine to delay further processing of the received content and adjust an operating parameter or parameters each associated with a corresponding event of the one or more events, the operating parameter or parameters track any combination of (i) a number of Sleep request messages initiated during processing of the received content by the virtual machine, or (ii) a cumulative Sleep time requested during processing of the received content by the virtual machine, or (iii) a number of calls initiated during processing of the received content by the virtual machine;comparison logic to compare the operating parameter or parameters to a threshold associated with the corresponding event, wherein the received content is classified as including delay-activated malware upon detecting that a value of the operating parameter or parameters exceed the threshold associated with the corresponding event; anda reporting module that, in response to classifying the received content as including delay-activated malware, issues an alert message for transmission from the system.

2. The system of claim 1, wherein the comparison logic including one or more comparators.

3. The system of claim 1, wherein the analysis logic comprises a sleep analysis logic that includes one or more counters, the sleep analysis logic, when executed by the one or more processors, monitors the number of Sleep request messages that are being repeated, each of the repeated Sleep request messages includes a parameter that indicates an amount of time requested for the corresponding Sleep request message.

4. The system of claim 3 further comprising time adjustment logic that is configured to operate in concert with the sleep analysis logic to compute a shortened Sleep time to accelerate malicious activities as detected by the comparison logic and the analysis logic to monitor the accelerated malicious activities as the one or more events.

5. The system of claim 1, wherein the analysis logic includes call analysis logic that includes one or more counters, the call analysis logic, when executed by the one or more processors, monitors for a presence of the repeated calls including Application Programming Interface (API) calls by accessing at least a counter of the one or more counters that maintains a count of a particular type of API call of the API calls initiated during processing of the received content within the virtual machine and alters a duration of a delay caused by the repeated calls by a shortened call response wait time for at least each of the particular type of API calls.

6. The system of claim 1 further comprising instruction pointer analysis logic that determines, during processing of the received content, whether an instruction pointer is repeatedly directed to a specific address or address range, wherein the instruction pointer being repeatedly directed to the specific address or address range operates as a criterion in classifying the received content as including delay-activated malware.

7. The system of claim 1 further comprising instruction pointer analysis logic that is configured to check, during processing of the content by the virtual machine, whether an instruction pointer has remained within an address range over a predetermined period of time, the detecting of the instruction pointer remaining within the address range over the predetermined period of time is a criterion for classifying the received content as including delay-activated malware.

8. The system of claim 1, wherein the analysis logic configured to monitor one or more events and adjust the operating parameter or parameters that track (i) the number of Sleep request messages initiated during processing of the received content by the virtual machine, and (ii) the cumulative Sleep time requested during processing of the received content by the virtual machine.

9. The system of claim 1, wherein the analysis logic corresponds to sleep analysis logic that monitors Sleep request messages and the system further comprises time adjustment logic that, when executed by the one or more processors, alters a duration of a delay caused by repeating Sleep request messages by shortening an amount of time allocated to each of the repeated Sleep request messages.

10. The system of claim 1, wherein the operating parameter or parameters further track at least one of (i) the number of Sleep request messages initiated during processing of the received content by the virtual machine, or (ii) the cumulative Sleep time requested during processing of the received content by the virtual machine, or (iii) the number of calls initiated during processing of the received content by the virtual machine, or (iv) a cumulative amount of time that called functions would need for execution,the comparison logic, when executed by the one or more processors, to compare (iv) the cumulative amount of time that the called functions would need for execution to a fourth threshold when the operating parameter or parameters is tracking the cumulative amount of time that the called functions would need for execution, andthe received content is classified as including delay-activated malware upon detecting that the cumulative amount of time that called functions would need for execution exceeds the fourth threshold.

11. The system of claim 1, wherein the comparison logic including one or more comparators to compare the operating parameter or parameters to the threshold by performing a comparison of a count value associated with the number of Sleep request messages to a first threshold when the one or more events correspond to the number of Sleep request messages initiated during processing by the virtual machine, wherein the received content is classified as including delay-activated malware upon detecting that the count value associated with the number of Sleep request messages exceeds the first threshold when the one or more events correspond to the number of Sleep request messages.

12. The system of claim 1, wherein the comparison logic including one or more comparators to compare the operating parameter or parameters to the threshold by performing a comparison of the cumulative Sleep time to a second threshold that is different than the first threshold when the one or more events correspond to the cumulative Sleep time, wherein the received content is classified as including delay-activated malware upon detecting that the cumulative Sleep time exceeds the second threshold when the one or more events correspond to the cumulative Sleep time.

13. The system of claim 1, wherein the comparison logic including one or more comparators to compare the operating parameter or parameters to the threshold by performing a comparison of a count value associated with the number of calls to a third threshold different than the first threshold and the second threshold when the one or more events correspond to the number of calls initiated during processing by the virtual machine, wherein the received content is classified as including delay-activated malware upon detecting that the count value associated with the number of calls exceeds the third threshold.

14. The system of claim 1, wherein the operating parameter or parameters track (i) the number of Sleep request messages initiated during processing of the received content by the virtual machine, (ii) the cumulative Sleep time requested during processing of the received content by the virtual machine, and (iii) the number of calls initiated during processing of the received content by the virtual machine.

15. A method for detecting time-bomb malware, comprising: instantiating one or more virtual machines that are adapted to receive content and process the received content for a period of time;monitoring for a presence of repeated calls produced during processing of the received content within the one or more virtual machines, the repeated calls to delay further processing of the received content;altering a duration of the delay that is caused by the repeated calls exceeding a prescribed threshold to accelerate operations by the received content to be conducted during the period of time;classifying the received content as including delay-activated malware upon detecting that the repeated calls exceeds the prescribed threshold; andresponsive to classifying the received content as including delay-activated malware, issuing an alert message indicating a presence of the delay-activated malware.

16. The method of claim 15, wherein the monitoring for the presence of the repeated calls includes monitoring for repeated calls and altering the duration of the delay by reducing a call response wait time for responding to each of the repeated calls.

17. The method of claim 15, wherein: the monitoring for the presence of the repeated calls includes monitoring for a plurality of Sleep request messages, each of the plurality of Sleep request messages includes a parameter that indicates an amount of time requested for a corresponding Sleep request message; andthe altering of the duration of the delay includes decreasing the amount of time requested for the corresponding Sleep request message.

18. The method of claim 15, wherein the monitoring for the presence of the repeated calls includes determining, during processing of the received content, whether an instruction pointer is repeatedly directed to a specific address or address range, the determining that the instruction pointer is repeatedly directed to the specific address or the address range operates as a criterion in classifying the received content as including delay-activated malware.

19. A non-transitory storage medium including software that, when executed by one or more processors, cause the software to perform operations comprising: monitoring for a presence of repeated calls produced during processing of content within a sandboxed environment, the repeated calls to delay further processing of the received content;altering a duration of the delay that is caused by the repeated calls exceeding a prescribed threshold to accelerate operations by the received content to be conducted during the period of time;classifying the received content as including delay-activated malware upon detecting that the repeated calls exceeds the prescribed threshold; andresponsive to classifying the received content as including delay-activated malware, issuing an alert message indicating a presence of the delay-activated malware.

20. The non-transitory storage medium of claim 19, wherein the repeated calls include Application Programming Interface (API) calls.

21. The non-transitory storage medium of claim 20, wherein the monitoring for the presence of the API calls performed by the software executed by the one or more processors comprises determining a count of a particular type of API call of the API calls initiated during processing of the received content within the sandboxed environment including a virtual machine.

22. The non-transitory storage medium of claim 21, wherein the altering of the duration of the delay performed by the software executed by the one or more processors comprises shortening call response wait time for at least each of the particular type of API calls.

23. The non-transitory storage medium of claim 19, wherein the repeated calls includes repeated Sleep calls.

24. The non-transitory storage medium of claim 19, wherein the altering the duration of the delay performed by the software executed by the one or more processors comprises reducing a call response wait time for responding to each of the repeated calls.

25. The non-transitory storage medium of claim 19, wherein the monitoring for the presence of the repeated calls performed by the software executed by the one or more processors comprises monitoring for a plurality of Sleep request messages, each of the plurality of Sleep request messages includes a parameter that indicates an amount of time requested for a corresponding Sleep request message.

26. The non-transitory storage medium of claim 25, wherein the altering of the duration of the delay performed by the software executed by the one or more processors comprises decreasing the amount of time requested for the corresponding Sleep request message.

27. The non-transitory storage medium of claim 19, wherein the monitoring for the presence of the repeated calls performed by the software executed by the one or more processors comprises determining, during processing of the received content, whether an instruction pointer is repeatedly directed to a specific address or address range, the determining that the instruction pointer is repeatedly directed to the specific address or the address range operates as a criterion in classifying the received content as including delay-activated malware.

28. The non-transitory storage medium of claim 19, wherein the altering of the duration of the delay performed by the software executed by the one or more processors comprises adjusting one or more operating parameters associated with a number of the repeated calls produced.

29. The non-transitory storage medium of claim 28, wherein the repeated calls correspond to Sleep request messages.

30. The non-transitory storage medium of claim 19, wherein the altering of the duration of the delay performed by the software executed by the one or more processors comprises adjusting one or more operating parameters corresponding to a cumulative Sleep time requested during processing of the repeated calls by the sandboxed environment including a virtual machine.

31. The system of claim 1, wherein at least the analysis logic and the reporting module correspond to software stored within the non-transitory storage medium and executed by the one or more processors.

32. The method of claim 15, wherein the repeated calls includes repeated Sleep calls.