Privacy-preserving distributed visual data processing
This disclosure presents embodiments for performing privacy-preserving distributed computing. For example, a workload that includes a plurality of tasks may be identified, and a workload graph may be generated to represent the workload. In addition, a device connectivity graph with connectivity information for a plurality of processing devices may be identified. A privacy policy associated with the workload, along with privacy level information associated with the plurality of processing devices, may also be identified. A privacy constraint may then be identified based on the privacy policy and the privacy level information. A workload schedule that includes a mapping of the workload onto the plurality of processing devices may be determined based on the privacy constraint, the workload graph, and the device connectivity graph. The workload schedule may be sent to the plurality of processing devices, which may execute the workload based on the workload schedule.
1. An apparatus, comprising:
a processor to:
identify a workload comprising a plurality of tasks;
generate a workload graph based on the workload, wherein the workload graph comprises information associated with the plurality of tasks;
identify a device connectivity graph, wherein the device connectivity graph comprises device connectivity information associated with a plurality of processing devices;
identify a privacy policy associated with the workload;
identify privacy level information associated with the plurality of processing devices;
identify a privacy constraint based on the privacy policy and the privacy level information; and
determine a workload schedule, wherein the workload schedule comprises a mapping of the workload onto the plurality of processing devices, and wherein the workload schedule is determined based on the privacy constraint, the workload graph, and the device connectivity graph; and
a communication interface to send the workload schedule to the plurality of processing devices, wherein the workload is executed by the plurality of processing devices based on the workload schedule.
2. The apparatus of claim 1 , wherein the processor to determine the workload schedule is further to solve an integer linear programming model based on the privacy constraint.
3. The apparatus of claim 1 , wherein the plurality of tasks is associated with processing sensor data from one or more sensors.
4. The apparatus of claim 3 , wherein the one or more sensors comprise one or more of:
a camera;
an infrared sensor; or
a laser-based sensor.
5. The apparatus of claim 3 , wherein the sensor data comprises visual data.
6. The apparatus of claim 1 , wherein the workload graph further comprises information associated with a plurality of task dependencies among the plurality of tasks.
7. The apparatus of claim 6 , wherein the privacy policy comprises a plurality of privacy requirements associated with the plurality of task dependencies.
8. The apparatus of claim 1 , wherein the device connectivity information comprises information associated with a plurality of device connectivity links among the plurality of processing devices.
9. The apparatus of claim 8 , wherein the privacy level information comprises a plurality of privacy levels associated with the plurality of device connectivity links.
10. A system, comprising:
a plurality of sensors to capture sensor data associated with an environment;
a plurality of processing devices, wherein the plurality of processing devices comprises a plurality of edge processing devices and a plurality of cloud processing devices, and wherein the plurality of processing devices is to:
identify a workload, wherein the workload comprises a plurality of tasks associated with processing the sensor data captured by the plurality of sensors;
generate a workload graph based on the workload, wherein the workload graph comprises information associated with the plurality of tasks;
identify a device connectivity graph, wherein the device connectivity graph comprises device connectivity information associated with the plurality of processing devices;
identify a privacy policy associated with the workload;
identify privacy level information associated with the plurality of processing devices;
identify a privacy constraint based on the privacy policy and the privacy level information;
determine a workload schedule, wherein the workload schedule comprises a mapping of the workload onto the plurality of processing devices, and wherein the workload schedule is determined based on the privacy constraint, the workload graph, and the device connectivity graph; and
distribute the workload schedule to the plurality of processing devices, wherein the workload is executed by the plurality of processing devices based on the workload schedule.
11. The system of claim 10 , wherein the plurality of processing devices to determine the workload schedule is further to solve an integer linear programming model based on the privacy constraint.
12. The system of claim 10 , wherein the plurality of sensors comprises one or more of:
a camera;
an infrared sensor; or
a laser-based sensor.
13. The system of claim 10 , wherein the workload graph further comprises information associated with a plurality of task dependencies among the plurality of tasks.
14. The system of claim 13 , wherein the privacy policy comprises a plurality of privacy requirements associated with the plurality of task dependencies.
15. The system of claim 10 , wherein the device connectivity information comprises information associated with a plurality of device connectivity links among the plurality of processing devices.
16. The system of claim 15 , wherein the privacy level information comprises a plurality of privacy levels associated with the plurality of device connectivity links.
17. At least one non-transitory machine accessible storage medium having instructions stored thereon, wherein the instructions, when executed on a machine, cause the machine to:
identify a workload comprising a plurality of tasks;
generate a workload graph based on the workload, wherein the workload graph comprises information associated with the plurality of tasks;
identify a device connectivity graph, wherein the device connectivity graph comprises device connectivity information associated with a plurality of processing devices;
identify a privacy policy associated with the workload;
identify privacy level information associated with the plurality of processing devices;
identify a privacy constraint based on the privacy policy and the privacy level information;
determine a workload schedule, wherein the workload schedule comprises a mapping of the workload onto the plurality of processing devices, and wherein the workload schedule is determined based on the privacy constraint, the workload graph, and the device connectivity graph; and
distribute the workload schedule to the plurality of processing devices, wherein the workload is executed by the plurality of processing devices based on the workload schedule.
18. The at least one non-transitory machine accessible storage medium of claim 17 , wherein the instructions that cause the machine to determine the workload schedule further cause the machine to solve an integer linear programming model based on the privacy constraint.
19. The at least one non-transitory machine accessible storage medium of claim 17 , wherein the plurality of tasks is associated with processing sensor data from one or more sensors.
20. The at least one non-transitory machine accessible storage medium of claim 17 , wherein:
the workload graph further comprises information associated with a plurality of task dependencies among the plurality of tasks; and
the privacy policy comprises a plurality of privacy requirements associated with the plurality of task dependencies.
21. The at least one non-transitory machine accessible storage medium of claim 17 , wherein:
the device connectivity information comprises information associated with a plurality of device connectivity links among the plurality of processing devices; and
the privacy level information comprises a plurality of privacy levels associated with the plurality of device connectivity links.
22. A method, comprising:
identifying a workload, wherein the workload comprises a plurality of tasks associated with processing sensor data from one or more sensors;
generating a workload graph based on the workload, wherein the workload graph comprises information associated with the plurality of tasks;
identifying a device connectivity graph, wherein the device connectivity graph comprises device connectivity information associated with a plurality of processing devices;
identifying a privacy policy associated with the workload;
identifying privacy level information associated with the plurality of processing devices;
identifying a privacy constraint based on the privacy policy and the privacy level information;
determining a workload schedule, wherein the workload schedule comprises a mapping of the workload onto the plurality of processing devices, and wherein the workload schedule is determined based on the privacy constraint, the workload graph, and the device connectivity graph; and
distributing the workload schedule to the plurality of processing devices, wherein the workload is executed by the plurality of processing devices based on the workload schedule.
23. The method of claim 22 , wherein determining the workload schedule comprises solving an integer linear programming model based on the privacy constraint.
24. The method of claim 22 , wherein:
the workload graph further comprises information associated with a plurality of task dependencies among the plurality of tasks; and
the privacy policy comprises a plurality of privacy requirements associated with the plurality of task dependencies.
25. The method of claim 22 , wherein:
the device connectivity information comprises information associated with a plurality of device connectivity links among the plurality of processing devices; and
the privacy level information comprises a plurality of privacy levels associated with the plurality of device connectivity links.