Autoscreenshot systems and methods for virtual operating system states
Described screenshot verification systems and methods for automatically verifying the integrity of a backup image or other process-of-interest using a screenshot verification system, as well as disaster recovery systems including said systems and performing said methods. In accordance with various aspects of the present disclosure, a virtual machine is booted and screenshots of the boot process are taken, which are used by a trained model, such as a convolutional neural network, to determine a boot state consistency. The systems and methods described deliver over 99% accuracy and do not involve regular expression analysis typical of conventional methods.
1. A method of evaluating integrity of a process-of-interest associated with a protected machine, the method comprising:
initiating, via an operating system managing the protected machine, the process-of-interest;
taking, via the operating system managing the protected machine, a plurality of screenshots of a display output while the process-of-interest is being initiated;
determining a state classification for each screenshot of the plurality of screenshots of the display output using a trained model, based on one or more determined attributes associated with each screenshot of the plurality of screenshots, wherein the trained model is an artificial neural network;
adding each screenshot of the plurality of screenshots of the display output and the state classification associated with each screenshot to a memory array;
testing at least a first portion of the memory array to determine a consistency class for the process-of-interest based on screenshots and the associated state classifications stored in at least the first portion of the memory array, wherein the consistency class represents a steady state classification over a particular period of time; and
performing one or more operations based on result of the testing.
2. The method of claim 1 , wherein the one or more operations include:
in response to the consistency class not being determined, collecting one or more additional screenshots;
adding the one or more additional screenshots to a second portion of the memory array; and
testing at least the second portion of the memory array to determine the consistency class for the process-of-interest.
3. The method of claim 1 , wherein the one or more operations include generating a report concerning the consistency class determined in connection with at least the first portion of the memory array, the report comprising a representative screenshot selected from the screenshots stored within the memory array and the consistency class.
4. The method of claim 1 , wherein the associated state classifications determined in association with each screenshot of the plurality of screenshots is at least one of: positive; informational; neutral; and negative.
5. The method of claim 4 , wherein at least the first portion of the memory array has a sample period m 1 and is tested to determine the consistency class by determining whether the first portion of the memory array has a common state classification associated with the one or more screenshots of the plurality of screenshots of the display output across the first portion of the memory array, wherein m 1 is an integer between 2 and 1000.
6. A method of evaluating integrity of at least a first backup image associated with a protected machine, the method comprising:
booting, via an operating system managing a first device, a virtual machine based on at least the first backup image;
taking, via the operating system managing the first device, a plurality of samples of display output of the virtual machine over a first duration;
determining, via a trained model of a prediction module, state classifications for one or more samples of the plurality of samples, based on one or more determined attributes associated with the one or more samples of the plurality of samples, wherein the trained model is an artificial neural network trained using a plurality of training images from one or more operating systems;
adding each sample of the plurality of samples and each associated state classifications to a first portion of a memory array; and
determining whether a consistency class is reached based on the plurality of samples added in the first portion of the memory array and the state classifications stored in the memory array, wherein the consistency class represents a steady state classification over a particular period of time.
7. The method of claim 6 , wherein the prediction module is located on a second device, the plurality of samples being communicated from the first device to the second device.
8. The method of claim 7 , the method further comprising:
generating a recommended action to be taken based on the consistency class; and
taking the recommended action, wherein the recommended action is taken by at least one of the first device and the second device.
9. The method of claim 8 , wherein the recommended action includes at least one of: reporting information regarding the consistency class; recording information regarding the consistency class; and ending the method.
10. The method of claim 7 , wherein the first device is the protected machine, and the second device is a screenshot verification server located remotely from the protected machine.
11. The method of claim 7 , wherein the first device is a remote storage server that is storing a backup image associated with the protected machine, and the second device is a screenshot verification server located remotely from the protected machine and the remote storage server.
12. The method of claim 7 , wherein the first device and the second device are the same device and are located remotely from the protected machine.
13. The method of claim 6 , wherein the state classification determined for each sample of the plurality of samples is at least one of: positive; informational; neutral; and negative.
14. The method of claim 6 , further comprising:
in response to determining that a consistent class is not reached:
collecting one or more additional screenshots;
adding the one or more additional screenshots to a second portion of the memory array;
testing at least the second portion of the memory array to determine the consistency class; and
repeating steps above one or more times until the consistency class is determined or a timeout duration is reached.
15. The method of claim 14 , wherein at least the first portion of the memory array has a sample period m 1 and is tested to determine the consistency class by determining whether the first portion of the memory array has a common state classification associated with the one or more samples of the plurality of samples of the virtual machine across the first portion of the memory array, wherein m 1 is an integer between 2 and 1000.
16. The method of claim 15 , wherein the second portion of the memory array has a sample period m 2 and is tested to determine the consistency class based on the second portion of the memory array by determining whether the second portion of the memory array has a common state classification associated with the one or more samples of the plurality of samples of the virtual machine stored in the memory array across the sample period m 2 , wherein m 2 is an integer between 2 and 1000.
17. The method of claim 16 , wherein at least the second portion of the memory array includes one or more samples and one or more state classifications not included in the first portion of the memory array.
18. The method of claim 6 , wherein the consistency class is at least one of: consistent positive; consistent negative; consistent neutral; and consistent informational.
19. A backup system comprising: a first device including at least one processor and memory including instructions that, when executed by the at least one processor, cause the first device to perform one or more of the following steps:
(a) obtaining a plurality of samples of display output of a virtual machine, booted based on at least a first backup image, over a first duration;
(b) communicating the plurality of samples to a prediction module comprising a trained model, wherein the trained model is an artificial neural network trained using a plurality of boot-state training images from one or more operating systems;
(c) determining, via the prediction module, state classifications for each sample of the plurality of samples of the virtual machine, based on one or more determined attributes associated with the one or more samples of the plurality of samples;
(d) adding each sample of the plurality of samples of the virtual machine and the state classifications determined in association with each sample to a memory array;
(e) testing at least a first portion of the memory array to determine a consistency class based on the plurality of samples and the state classifications stored in at least the first portion of the memory array, wherein the consistency class represents a steady state classification over a particular period of time; and
performing one or more operations based on the consistency class.
20. The backup system of claim 19 , wherein the operations include:
repeating steps (b) through (e) and testing at least a second portion of the memory array to determine a consistency class based on the plurality of samples and state classifications stored in at least the second portion of the memory array, wherein the second portion of the memory array includes one or more samples and one or more state classifications not included in the first portion of the memory array;
generating a report concerning the consistency class determined in connection with at least the first portion of the memory array, the report comprising a representative sample selected from the plurality of samples stored within the memory array and the consistency class; and
generating a recommended action to be taken based on the consistency class and taking the recommended action, wherein the recommended action including at least one of: reporting information regarding the consistency class; recording information regarding the consistency class; and terminating operations.