Automatic and dynamic performance benchmarking and scoring of applications based on crowdsourced traffic data
The disclosure describes various aspects of crowdsourcing traffic data for automatic and dynamic benchmarking of applications. In an aspect, an intelligence layer, communicatively coupled to a data collection layer and a visualization layer, is configured to receive traffic data from data sources (e.g., physical appliances, probes) in the data collection layer, the data sources being associated with multiple customers, and the traffic data being associated with at least one application (e.g., word processing, video streaming) used by the multiple customers. The intelligence layer is a cloud-based layer further configured to process the traffic data to determine performance thresholds for the at least one application, and may send one or more of the performance thresholds to a data source for a different customer to be used for benchmarking the at least one application for the different customer.
1. A system comprising:
a processor;
a network interface; and
a machine-readable medium having program code executable by the processor to cause the system to,
obtain crowdsourced traffic data from distributed data sources via the network interface, wherein the crowdsourced traffic data comprise data packets collected from network sites of multiple customers and generated from applications being used by users of the multiple customers;
determine a set of one or more performance thresholds for a first of the applications based, at least in part, on information from the crowdsourced traffic data; and
score current performance of the first application at each network site of first customer with respect to the set of performance thresholds, wherein the first customer is different than the multiple customers.
2. The system of claim 1 , wherein the distributed data sources comprise probes distributed across devices of the multiple customers.
3. The system of claim 1 , wherein the performance thresholds include time response thresholds for at least the first application.
4. The system of claim 1 , wherein the performance thresholds indicate a minimum acceptable value of a performance metric associated with the first application.
5. The system of claim 1 , wherein the program code executable to score performance of the first application comprises program code executable by the processor to cause the system to produce, for each network site of the first customer, an index value based, at least in part, on the set of performance thresholds.
6. The system of claim 5 , wherein the index value ranges from 0 to N, N being an integer number, and wherein the first application has a better performance for the different customer as benchmarked against the set of one or more performance thresholds when the index value is closer to N than when it is closer to 0.
7. The system of claim 1 , wherein the distributed data sources include one or more of a public cloud data source, a software-defined network (SDN) or a network as a service (NaaS) data source, a virtual network function (VNF) on universal customer premises equipment (uCPE) data source, a virtual appliance, or a physical appliance.
8. The system of claim 1 , wherein the machine-readable medium further has program code executable by the processor to cause the system to:
segment the crowdsourced traffic data into a plurality of segments of crowdsourced traffic data,
wherein the program code to determine the set of one or more performance thresholds for the first application comprises the program code executable by the processor to cause the system to determine, from each segment of crowdsourced traffic data, at least one performance threshold for the first application,
wherein the program code to score performance of the first application for each network site of the first customer comprises program code executable by the processor to cause the system to score performance of the first application for each network site of the first customer with respect to a first of the set of performance thresholds that was determined from a first segment of the crowdsourced traffic data based on selection of the first segment for the first customer.
9. The system of claim 8 wherein the plurality of segments corresponds to at least one of industry segments and geographic segments.
10. The system of claim 1 , wherein the machine-readable medium further has program code that is executable by the processor to cause the system to:
obtain additional crowdsourced traffic data that includes application performance information for the first application; and
update the set of one or more performance thresholds based, at least in part, on information from the additional crowdsourced traffic data.
11. A method comprising:
obtaining crowdsourced traffic data from data sources across network sites of a plurality of customers, wherein the crowdsourced traffic data includes information about application performance for an application used by users of the plurality of customers;
determining a set of one or more performance thresholds for the application based, at least in part, on information from the crowdsourced traffic data; and
scoring current performance of the application at each network site of a first customer with respect to the set of one or more performance thresholds.
12. The method of claim 11 , wherein the data sources comprise probes distributed across devices of the plurality of customers.
13. The method of claim 11 , wherein the performance thresholds indicate a minimum acceptable value of a performance metric associated with the application.
14. The method of claim 11 , wherein scoring current performance of the application for the first customer comprises producing an index value for each network site of the first customer based, at least in part, on the set of performance thresholds.
15. The method of claim 14 , wherein the index value ranges from 0 to N, N being an integer number, and wherein the application has a better performance for the first customer as benchmarked against the set of one or more performance thresholds when the index value is closer to N than when it is closer to 0.
16. The method of claim 11 , further comprising:
segmenting the crowdsourced traffic data into a plurality of segments of crowdsourced traffic data,
wherein determining the set of one or more performance thresholds for the application comprises determining, from each segment of crowdsourced traffic data, at least one performance threshold for the application,
wherein scoring performance of the application for the first customer comprises scoring performance of the application with respect to a first of the set of performance thresholds that was determined from a first segment of the crowdsourced traffic data based on selection of the first segment for the first customer.
17. The method of claim 16 wherein the plurality of segments corresponds to at least one of industry segments and geographic segments.
18. The method of claim 11 , further comprising:
obtaining additional crowdsourced traffic data that includes application performance information for the application; and
updating the set of one or more performance thresholds based, at least in part, on information from the additional crowdsourced traffic data.
19. The method of claim 11 , wherein the plurality of customers includes the first customer.
20. A non-transitory machine-readable medium having program code stored thereon, the program code to:
obtain crowdsourced traffic data from data sources across network sites of a plurality of customers, wherein the crowdsourced traffic data includes information about application performance for an application used by users of the plurality of customers;
determine a set of one or more performance thresholds for the application based, at least in part, on information from the crowdsourced traffic data; and
score current performance of the application at each network site of a first customer with respect to the set of performance thresholds.
21. The non-transitory machine-readable medium of claim 20 , wherein the data sources comprise probes distributed across devices of the plurality of customers.
22. The non-transitory machine-readable medium of claim 20 , wherein the performance thresholds indicate a minimum acceptable value of a performance metric associated with the application.
23. The non-transitory machine-readable medium of claim 20 , wherein the program code to score current performance of the application for a first customer comprises program code to produce an index value for each network site of the first customer based, at least in part, on the set of performance thresholds.
24. The non-transitory machine-readable medium of claim 23 , wherein the index value ranges from 0 to N, N being an integer number, and wherein the application has a better performance for the first customer as benchmarked against the set of one or more performance thresholds when the index value is closer to N than when it is closer to 0.
25. The non-transitory machine-readable medium of claim 20 , further comprising program code to:
segment the crowdsourced traffic data into a plurality of segments of crowdsourced traffic data,
wherein the program code to determine the set of one or more performance thresholds for the application comprises program code to determine, from each segment of crowdsourced traffic data, at least one performance threshold for the application,
wherein the program code to score performance of the application at each network site of the first customer comprises program code to score performance of the application at each network site with respect to a first of the set of performance thresholds that was determined from a first segment of the crowdsourced traffic data based on selection of the first segment for the first customer.
26. The non-transitory machine-readable medium of claim 25 wherein the plurality of segments corresponds to at least one of industry segments and geographic segments.
27. The non-transitory machine-readable medium of claim 20 further comprising program code to:
obtain additional crowdsourced traffic data that includes application performance information for the application; and
update the set of one or more performance thresholds based, at least in part, on information from the additional crowdsourced traffic data.
28. The non-transitory machine-readable medium of claim 20 , wherein the plurality of customers includes the first customer.