HISTORY-BASED CLASSIFICATION OF TRAFFIC INTO QOS CLASS WITH SELF-UPDATE

• US 20180131620A1
• Filed: 11/10/2016
• Published: 05/10/2018
• Est. Priority Date: 11/10/2016
• Status: Active Grant

• Current Assignee: Hughes Network Systems LLC (Echostar Corporation)
• Original Assignee: Hughes Network Systems LLC (Echostar Corporation)
• Inventors: Su, Chi-Jiun, Jain, Kaustubh, Hong, Se Gi
• Granted Patent: US 10,341,241 B2
• Time in Patent Office: Days
• CPC Class Codes: H04L 47/2441 relying on flow classification, e.g. using integrated services [IntServ]
• CPC Class Codes: H04L 47/2475 for supporting traffic characterised by the type of applications
• CPC Class Codes: H04L 47/2483 involving identification of individual flows
• CPC Class Codes: H04L 61/45 Network directories; Name-to-address mapping
• CPC Class Codes: H04L 69/16 Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]

Abstract

Methods and systems of history-based instantaneous QoS class classification of traffic flow. QoS classifications are mapped to target destinations such that, when a first flow parcket packet of a traffic flow requesting access to a resource associated with the target destination, a classifier can instantaneously classify the traffic flow as belonging to a particular QoS class for servicing of the traffic flow.

Claims

1. A method of traffic classification, comprising:

receiving, at a traffic classifier, a first flow packet of a current traffic flow;

identifying, by the traffic classifier, a target destination of the current traffic flow based on the first flow packet;

checking a history, by the traffic classifier, wherein the history associates a plurality of target destinations with an associated classification type; and

tagging, by the traffic classifier, the current traffic flow as belonging to the classification type associated with the target destination;

wherein the classification type comprises one of a plurality of quality of service (“

QoS”

) classes having one or more QoS requirements.

2. The method of claim 1, wherein the first flow packet is a TCP SYN packet, and identifying a target destination further comprises identifying a target destination of the current traffic flow based on the TCP SYN packet and a DNS response message.

3. The method of claim 1, wherein the traffic classifier is implemented within one or more of:

a consumer premise equipment (“

CPE”

) device;

a network device located at a network gateway;

or a network device located at a bottleneck link within a network.

4. The method of claim 1, wherein the target destination is identified by a fully qualified domain name (“

FQDN”

).

5. The method of claim 4, wherein the history comprises a first lookup table mapping one or more IP addresses to one or more FQDNs, and a second lookup table mapping each FQDN of the one or more FQDNs to a classification type.

6. The method of claim 1, wherein the target destination is identified by an IP address.

7. The method of claim 6, wherein the history comprises a lookup table mapping each IP address of a plurality of IP addresses to a classification type.

8. The method of claim 1, wherein the current traffic flow comprises encrypted traffic.

9. The method of claim 1, wherein checking the history comprises if the history does not contain an association between the target destination and a classification type, determining a recommended classification type based on information in the history.

10. The method of claim 9, wherein determining a recommended classification type comprises using a recommendation scheme, the recommendation scheme comprising one of:

a similarity-based scheme;

a keyword-based scheme;

or a popularity-based scheme.

11. The method of claim 10, wherein the traffic classifier is configured to use two or more recommendation schemes.

12. The method of claim 1, further comprising:

collecting a plurality of traffic characteristics associated with the current traffic flow;

updating the history based on the collected plurality of traffic characteristics at completion of the current traffic flow.

13. The method of claim 1, further comprising:

during the current traffic flow, performing real-time classification of the current traffic flow based on measure traffic characteristics; and

if a real-time classification type of the current traffic flow differs from the classification type tagged based on the history, identifying the real-time classification type.

14. An instantaneous traffic classification system, comprising:

a traffic classifier comprising a history comprising one or more lookup tables mapping a target destination to a classification type of one or more classification types; and

one or more queues configured to store a plurality of traffic flows, each queue associated with each classification type of the one or more classification types;

wherein the traffic classifier is configured to classify a current traffic flow as a classification type of the one or more classification types before any content packets of the current traffic flow are transmitted.

15. The system of claim 14, wherein the traffic classifier is implemented within one or more of:

a consumer premise equipment (“

CPE”

) device;

a network device located at a network gateway;

or a network device located at a bottleneck link within a network.

16. The system of claim 14, wherein the target destination is identified by a fully qualified domain name (“

FQDN”

).

17. The system of claim 16, wherein the history comprises a first lookup table mapping one or more IP addresses to one or more FQDNs, and a second lookup table mapping each FQDN of the one or more FQDNs to a classification type.

18. A method of updating a history for a history-based traffic classification system, comprising:

determining a flow information set for a target destination of a plurality of target destinations, wherein the flow information set includes a flow object for each traffic flow associated with each target destination, each flow object identifying a flow classification type and one or more weights;

determining a total weight associated with each flow classification type present in the flow information set;

determining an overall total weight comprising the sum of all weights present in the flow information set;

determining normalized weights for each flow classification type present in the flow information set;

determining a maximum normalized weight from a set of normalized weights;

compare the maximum normalized weight against a threshold value;

if the maximum normalized weight exceeds the threshold value, setting the flow classification type associated with the maximum normalized weight as a classification type for the target destination; and

updating a history to identify the target destination as associated with the set classification type for the target destination.

19. The method of claim 18, further comprising:

if the maximum normalized weight does not exceed the threshold, determining a preferred classification type between a set of flow classification types present in the flow information set; and

;

update the history to identify the target destination as associated with the preferred classification type.

20. The method of claim 19, determining the preferred classification type comprises checking a Top-N table, the Top-N table having entries identifying the preferred classification types when N classification types are present in the flow information set.

21. The method of claim 18, wherein the target destination is a Fully Qualified Domain Name (“

FQDN”

).

22. The method of claim 18, wherein the target destination is an IP address.