Granted Patent B2

US 9,848,009 · App. 15/465,623 · Granted Dec 19, 2017

Inventors: Avi Turgeman (Cambridge, MA); Itai Novick (Rehovot, IL)

Assignee: BioCatch Ltd.

H04L63/1425H04L63/083H04L63/145

Abstract

Claims (87)

1. A method comprising:

(A) detecting an automated malware that emulates human interactions with a computerized service;

wherein the detecting of step (A) comprises:

(a) monitoring input-unit interactions of an electronic device that is utilized by a user to interact with said computerized service;

(b) analyzing said input-unit interactions;

(c) determining that it is humanly-impossible for a human to perform said input-user interactions;

(d) based on the determining of step (c), determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user;

wherein the method comprises:

injecting an input-output aberration into a web-page, and monitoring whether manual corrective actions were manually performed in response to the input-output aberration;

generating a list of usage sessions in which at least one of the following conditions holds true: (I) no corrective actions were performed, or (II) corrective actions that were performed were insufficient to adequately cure the input-output aberration;

diluting said list of usage sessions, by removing usage sessions that do not exhibit a mismatch between (I) input-unit events as actually registered by one or more input units of the electronic device, and (II) content that was registered by the electronic device as allegedly being manually entered.

2. The method of claim 1 , wherein steps (b) and (c) and (d) comprise:

detecting that said input-unit interactions comprise an on-screen displacement-jump of an on-screen pointer, from a first on-screen location, to a second and non-adjacent on-screen location;

based on said on-screen displacement-jump of the on-screen pointer, determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user.

3. The method of claim 1 , wherein steps (b) and (c) and (d) comprise:

detecting that said input-unit interactions comprise an on-screen displacement-jump of an on-screen pointer, from a first on-screen location, to a second and non-adjacent on-screen location; and further detecting that said on-screen displacement-jump is accompanied by one or more input-user gestures that are insufficient for commanding said displacement-jump;

based on said on-screen displacement-jump of the on-screen pointer, determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user.

4. The method of claim 1 , wherein steps (b) and (c) and (d) comprise:

(i) detecting that said input-unit interactions exhibit rapid displacement-jumps of an on-screen pointer, among non-adjacent on-screen locations, wherein said rapid displacement-jumps comprise two or more on-screen pointer displacement-jumps that occur within a pre-defined time period;

(ii) based on step (i), determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user.

5. The method of claim 1 , wherein steps (b) and (c) and (d) comprise:

(i) detecting a first click or tap, performed at a first on-screen location, occurred concurrently with a second click or tap performed at a second, different, on-screen location;

(ii) based on step (i), determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user.

6. The method of claim 1 , wherein steps (b) and (c) and (d) comprise:

(i) monitoring key-down events, and monitoring key-up events, during a usage session in which said electronic device exhibits reception of keyboard input;

(ii) determining that the order of the key-down events and the key-up events, during said usage session, does not match an expected order of key-down events and key-up events that is expected to be observed if an input unit of said electronic device is utilized for typing by a human user;

(iii) based on step (ii), determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user.

7. The method of claim 1 , wherein steps (b) and (c) and (d) comprise:

(i) monitoring key-down events, and monitoring key-up events, during a usage session in which said electronic device exhibits reception of keyboard input;

(ii) determining that (I) a timing scheme that characterizes the monitored key-down events and the monitored key-up events, during said usage session, does not match (II) a timing scheme of entry of typed characters as actually registered by the electronic device;

(iii) based on the determining of step (i), determining that said electronic device is necessarily controlled by an automated attacking module, and not by a legitimate human user.

8. The method of claim 1 , wherein steps (b) and (c) and (d) comprise:

(i) continuously monitoring mouse events, during a usage session in which said electronic device exhibits reception of mouse-based input;

(ii) determining that during a first period of time within said usage session, the monitored mouse events exhibit a first sampling rate;

(iii) determining that during a second period of time within said usage session, the monitored mouse events exhibit a second, different, sampling rate;

(iv) based on steps (ii) and (iii), determining that said electronic device is necessarily controlled by an automated attacking module, and not by a legitimate human user.

9. The method of claim 1 , wherein steps (b) and (c) and (d) comprise:

(i) detecting that an input-unit level of the electronic device reports that a message of M characters were manually entered via an input-unit of the electronic device;

(ii) detecting that said electronic device sends to a remote server, an outgoing message of N characters that was allegedly typed on said electronic device, wherein N is different than M;

(iii) based on the determining of steps (i) and (ii), further determining that said electronic device is necessarily controlled by an automated attacking module, and not by a legitimate human user.

10. The method of claim 1 , comprising:

(i) detecting that the electronic device sends to a remote server, a message indicating that a string of characters was manually entered by a user in a particular field of a particular fillable form;

(ii) based on monitoring of input-unit interactions of said electronic device, determining that the input unit of the electronic device registered manual input during the time-period that corresponds to alleged manual data entry of said particular field of said particular fillable form; and further determining that the manual input, as registered by the input unit of the electronic device, does not yield said string of characters that said message indicates as manually entered;

(iii) based on steps (i) and (ii), further determining that said electronic device is necessarily controlled by an automated attacking module, and not by a legitimate human user.

11. A method comprising:

(A) detecting an automated malware that emulates human interactions with a computerized service;

wherein the detecting of step (A) comprises:

(a) monitoring input-unit interactions of an electronic device that is utilized by a user to interact with said computerized service;

(b) analyzing said input-unit interactions;

(c) determining that it is humanly-impossible for a human to perform said input-user interactions;

(d) based on the determining of step (c), determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user;

wherein the method comprises:

generating a list of usage sessions that exhibit a mismatch between (I) input-unit events as actually registered by one or more input units of the electronic device, and (II) content that was registered by the electronic device as allegedly being manually entered;

diluting said list of usage sessions by: (a) injecting an input-output aberration into a web-page, and monitoring whether manual corrective actions were manually performed in response to the input-output aberration; and (b) detecting that at least one of the following conditions holds true: (I) no corrective actions were performed, or (II) corrective actions that were performed were insufficient to adequately cure the input-output aberration.

12. The method of claim 11 , wherein steps (b) and (c) and (d) comprise:

detecting that said input-unit interactions comprise an on-screen displacement-jump of an on-screen pointer, from a first on-screen location, to a second and non-adjacent on-screen location; and further detecting that said on-screen displacement-jump is not accompanied by any input-user gesture that is required for commanding said displacement-jump;

based on said on-screen displacement-jump of the on-screen pointer, determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user.

13. The method of claim 11 , wherein steps (b) and (c) and (d) comprise:

(i) detecting that said input-unit interactions comprise an on-screen displacement-jump of an on-screen pointer, from a first on-screen location, to a second and non-adjacent on-screen location; and further determining that said on-screen displacement-jump of the on-screen pointer does not match statistical distribution of input-unit utilization by human users;

(ii) based on step (i), determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user.

14. The method of claim 11 , wherein steps (b) and (c) and (d) comprise:

(i) detecting a first click or tap, performed at a first on-screen location;

(ii) subsequently, detecting a second click or tap, performed at a second on-screen location;

(iii) detecting that no input-unit events were registered between (I) detection of the first click or tap at the first on-screen location, and (II) detection of the second click or tap at the second on-screen location;

(iv) based on step (iii), determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user.

15. The method of claim 11 , wherein steps (b) and (c) and (d) comprise:

(i) monitoring key-down events, and key-up events, during a usage session in which said electronic device exhibits reception of keyboard input;

(ii) determining that the number of key-down events does not match the number of key-up events, during said usage session;

(iii) based on step (ii), determining that said input-unit interactions were necessarily performed by said automated script that emulates human interactions, and not by a human user.

16. The method of claim 11 , wherein steps (b) and (c) and (d) comprise:

(i) monitoring key-down events, and monitoring key-up events, during a usage session in which said electronic device exhibits reception of keyboard input;

(ii) determining that at least one of: (I) the key-down events that were registered by the electronic device during said usage session, and (II) the key-up events that were registered by the electronic device during said usage session, does not match the number of characters that said electronic device registered as received from a keyboard;

(iii) based on the determining of step (i), determining that said electronic device is necessarily controlled by an automated attacking module, and not by a legitimate human user.

17. The method of claim 11 , wherein steps (b) and (c) and (d) comprise:

(i) continuously monitoring keyboard events, during a usage session in which said electronic device exhibits reception of keyboard-based input;

(ii) determining that during a first period of time within said usage session, the monitored keyboard events exhibit a first sampling rate;

(iii) determining that during a second period of time within said usage session, the monitored keyboard events exhibit a second, different, sampling rate;

(iv) based on steps (ii) and (iii), determining that said electronic device is necessarily controlled by an automated attacking module, and not by a legitimate human user.

18. The method of claim 11 , wherein steps (b) and (c) and (d) comprise:

(i) continuously monitoring keyboard events and mouse events, during a usage session in which said electronic device exhibits reception of keyboard-based input;

(ii) determining that during said usage session, the monitored keyboard events exhibit a first sampling rate;

(iii) determining that during said usage session, the monitored mouse events exhibit a second, different, sampling rate;

(iv) based on steps (ii) and (iii), determining that said electronic device is necessarily controlled by an automated attacking module, and not by a legitimate human user.

19. The method of claim 11 , wherein steps (b) and (c) and (d) comprise:

(i) detecting that the electronic device sends to a remote server, a message indicating that a string of characters was manually entered by a user in a particular field of a particular fillable form;

(ii) based on monitoring of input-unit interactions of said electronic device, determining that the input unit of the electronic device did not register any manual input during the time-period that corresponds to alleged manual data entry of said particular field of said particular fillable form;

(iii) based on the determining of step (ii), further determining that said electronic device is necessarily controlled by an automated attacking module, and not by a legitimate human user.

Assignments (4)

Continuity (11)