Automatic creation of Internet-of-Things device descriptors
The present technology automates the process of collecting and using data from IoT devices to improve IoT security and management while limiting the manual entry requirements of conventional IoT platforms. The technology receives data from IoT devices and determines behavioral characterizations of the IoT devices based on the data. The technology then uses a generative model to create a file representative of the behavioral characterizations of the IoT device, and uses that file to represent new IoT devices with similar attributes to those represented by the file. In doing so, the technology can improve security by, for example, creating a MUD file representative of a new IoT device with no associated MUD file. The technology can also create a file representative of the IoT device to improve onboarding of an IoT device in, for example, an IoT operations dashboard.
1 . A method comprising:
receiving data indicating attributes of a first internet-of-things device (first IoT device);
prompting a generative model to create a file based on the data, wherein the file is a Manufacturer Usage Description (MUD) file and specifies permitted communications and operational parameters of the first IoT device;
determining, by a network controller, that a second IoT device is a same type of device as the first IoT device by analyzing the second IoT device and concluding the second IoT device includes similar attributes as the first IoT device; and
associating the file with the second IoT device, including storing the MUD file on a MUD server and implementing a MUD policy based on the MUD file for the second IoT device by configuring network equipment to enforce the permitted communications.
2 . The method of claim 1 , wherein the data includes connectivity data, attribute data, and attribute access procedure data.
3 . The method of claim 1 , wherein receiving the data includes analyzing data packets from the first IoT device or the second IoT device with a Transmission Control Protocol dump (TCP dump).
4 . The method of claim 1 , wherein the data is a manual identifying attributes of the first IoT device.
5 . The method of claim 1 , further comprising uploading the file to an IoT operations dashboard to input the attributes of the first IoT device into the IoT operations dashboard.
6 . The method of claim 1 , further comprising training the generative model with training files that are a same format as the file.
7 . A network controller comprising:
a storage configured to store instructions; and
at least one processor configured to execute the instructions and cause the at least one processor to:
receive data indicating attributes of a first internet-of-things device (first IoT device);
prompt a generative model to create a file based on the data, wherein the file is a Manufacturer Usage Description (MUD) file and specifies permitted communications and operational parameters of the first IoT device;
determine that a second IoT device is a same type of device as the first IoT device by analyzing the second IoT device and concluding the second IoT device includes similar attributes as the first IoT device; and
associate the file with the second IT device, including storing the MUD file on a MUD server and implementing a MUD policy based on the MUD file for the second IoT device by configuring network equipment to enforce the permitted communications.
8 . The network controller of claim 7 , wherein the data includes connectivity data, attribute data, and attribute access procedure data.
9 . The network controller of claim 7 , wherein receiving the data includes analyzing data packets from the first IoT device or the second IoT device with a TCP dump.
10 . The network controller of claim 7 , wherein the data is a manual identifying attributes of a the first IoT device.
11 . The network controller of claim 7 , wherein the at least one processor is configured to execute the instructions and further cause the at least one processor to upload the file to an IoT operations dashboard to input the attributes of the first IoT device into the IoT operations dashboard.
12 . The network controller of claim 7 , wherein the at least one processor is configured to execute the instructions and further cause the at least one processor to train the generative model with training files that are a same format as the file.
13 . A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor, cause the at least one processor to:
receive data indicating attributes of a first internet-of-things device (first IoT device);
prompt a generative model to create a file based on the data, wherein the file is a Manufacturer Usage Description (MUD) file and specifies permitted communications and operational parameters of the first IoT device;
determine that a second IoT device is a same type of device as the first IoT device by analyzing the second IoT device and concluding the second IoT device includes similar attributes as the first IoT device; and
associate the file with the second IoT device, including storing the MUD file on a MUD server and implementing a MUD policy based on the MUD file for the second IoT device by configuring network equipment to enforce the permitted communications.
14 . The non-transitory computer-readable storage medium of claim 13 , wherein the data includes connectivity data, attribute data, and attribute access procedure data.
15 . The non-transitory computer-readable storage medium of claim 13 , wherein receiving the data includes analyzing data packets from the first IoT device or the second IoT device, respectively, with a TCP dump.
16 . The non-transitory computer-readable storage medium of claim 13 , wherein the data is a manual identifying attributes of the first IoT device.
17 . The non-transitory computer-readable storage medium of claim 13 , wherein the instructions further cause the at least one processor to upload the file to an IoT operations dashboard to input the attributes of the first IoT device into the IoT operations dashboard.