# US 2021/0344607 A1 · System and method for classifying network devices

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0344607 A1
- Document ID: US20210344607A1
- Status: Published application
- Kind: A1
- Application number: 17/244,336
- Filed: 2021-04-29
- Published: 2021-11-04
- Earliest claimed: 2020-04-30
- Applicants: Sandvine Corporation
- Inventors: Shyam SREEVALSAN; Rajeswara Rao MUTHYALA
- Web page: https://patentgrep.com/patent/US20210344607A1

## CPC

- H04L 47/2441: relying on flow classification, e.g. using integrated services [IntServ]
- G06N 20/00: Machine learning
- H04L 41/145: {involving simulating, designing, planning or modelling of a network}
- H04L 41/16: using machine learning or artificial intelligence
- H04L 41/5025: by proactively reacting to service quality change, e.g. by reconfiguration after service quality degradation or upgrade
- H04L 43/026: using flow identification
- H04L 47/2483: involving identification of individual flows
- Y02D 30/50: in wire-line communication networks, e.g. low power modes or reduced link rate
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- Y02D: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE

## Abstract

A method and system for classifying a device accessing a computer network. The method including: providing a framework of models configured to classify the device; reviewing a network traffic flow associated with a device; extracting flow attributes associated with a network traffic flow; deriving further flow attributes based on the extracted flow attributes; determining at least one model of the framework of models based on the derived flow attributes and extracted encrypted flow attributes; and classifying the device associated with the network traffic flow based on the at least one model. The system includes: a learning engine configured to provide a framework of models; a packet processing engine configured to review a network traffic flow associated with a device; a device classification engine configured to extract flow attributes, derive further flow attributes and determine at least one model; and a device information aggregator configured to classify the device.

## Claims

### Claim 1

A method for classifying a device accessing a computer network, the method comprising:
providing a framework of models configured to classify the device;
reviewing a network traffic flow associated with the device;
extracting flow attributes associated with the network traffic flow;
deriving further flow attributes based on the extracted flow attributes;
determining at least one model of the framework of models based on the derived flow attributes and extracted flow attributes; and
classifying the device associated with the network traffic flow based on the at least one model.

### Claim 2 (depends on 1)

A method for classifying a device according to claim 1, wherein classifying the device comprises determining a device type and a platform of the device.

### Claim 3 (depends on 2)

A method for classifying a device according to claim 2 wherein the device type is classified as one of the group comprising: a tablet, a phone, a laptop, a desktop, a home device and an Internet-of-things device.

### Claim 4 (depends on 2)

A method for classifying a device according to claim 2 wherein the device platform is classified as one of the group comprising: a Windows platform, an Android platform, and an iOS platform.

### Claim 5 (depends on 1)

A method for classifying a device according to claim 1, further comprising:
determining a confidence level of the classification of the device associated with the network traffic flow;
comparing the confidence level with a predetermined threshold;
if the confidence level is higher than the predetermined threshold, marking the device associated with the network traffic flow as classified; and
if the confidence level is lower than the predetermined threshold, determining whether there is User Agent string associated with the traffic flow;
if there is a User Agent string available, extracting device data from the User Agent String; and
updating the classification with the extracted device data.

### Claim 6 (depends on 1)

A method for classifying a device according to claim 1, further comprising:
determining an accuracy level for at least one model of the frameworks of models;
comparing the accuracy level to a predetermined accuracy threshold;
if the accuracy level is higher than the accuracy threshold, continue using the at least one model of the framework of models;
otherwise, updating or removing the model from the framework of models.

### Claim 7 (depends on 2)

A method for classifying a device according to claim 2, further comprising:
determining whether there is further device data to be classified; and
reviewing flow attributes associated with the further device data.

### Claim 8 (depends on 1)

A method for classifying a device according to claim 1, further comprising:
determining at least one traffic action associated with the device classification; and
applying the at least one traffic action to the network traffic flow.

### Claim 9

A system for classifying a device accessing a computer network, the system comprising:
a learning engine configured to provide a framework of models configured to classify the device;
a packet processing engine configured to review a network traffic flow associated with the device;
a device classification engine configured to extract flow attributes associated with the network traffic flow, derive further flow attributes based on the extracted flow attributes and determine at least one model of the framework of models based on the derived flow attributes and extracted encrypted flow attributes; and
a device information aggregator configured to classify the device associated with the network traffic flow based on the at least one model.

### Claim 10 (depends on 1)

A system for classifying a device according to claim 1, wherein the device classification engine comprises:
a device platform classifier configured to classify a device platform based on the extracted and derived flow attributes; and
a device type classifier configured to classify a device based on the extracted and derived flow attributes.

### Claim 11 (depends on 10)

A system for classifying a device according to claim 10 wherein the device type is classified as one of the group comprising: a tablet, a phone, a laptop, a desktop, a home device and an Internet-of-things device.

### Claim 12 (depends on 10)

A system for classifying a device according to claim 10 wherein the device platform is classified as one of the group comprising: a Windows platform, an Android platform, and an iOS platform.

### Claim 13 (depends on 9)

A system for classifying a device according to claim 9, wherein the device classification engine is further configured to:
determine a confidence level of the classification of the device associated with the network traffic flow;
compare the confidence level with a predetermined threshold;
if the confidence level is higher than the predetermined threshold, mark the device associated with the network traffic flow as classified; and
if the confidence level is lower than the predetermined threshold, determine whether there is User Agent string associated with the traffic flow;
if there is a User Agent string available, having a User Agent Parser extract device data from the User Agent String; and
wherein the device information aggregator is configured to update the classification with the extracted device data.

### Claim 14 (depends on 9)

A system for classifying a device according to claim 9, further comprising:
a model monitoring module configured to:
determine an accuracy level for at least one model of the frameworks of models;
compare the accuracy level to a predetermined accuracy threshold;
if the accuracy level is higher than the accuracy threshold, continue using the at least one model of the framework of models;
otherwise, update or remove the model from the framework of models.

### Claim 15 (depends on 10)

A system for classifying a device according to claim 10, wherein the device classification engine is configured to:
determine whether there is further device data to be classified; and
review flow attributes associated with the further device data.

### Claim 16 (depends on 9)

A system for classifying a device according to claim 9, wherein the system is further configured to:
determine at least one traffic action associated with the device classification; and
apply the at least one traffic action to the network traffic flow.
