US 12,155,548 B1
Anomaly detection in content delivery networks
Samrat Karak, Bothell, WA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Jul. 27, 2021, as Appl. No. 17/443,794.
Int. Cl. H04L 43/0823 (2022.01); G06N 20/00 (2019.01); H04L 9/40 (2022.01); H04L 41/16 (2022.01); H04L 43/028 (2022.01); H04L 43/06 (2022.01)
CPC H04L 43/0847 (2013.01) [G06N 20/00 (2019.01); H04L 41/16 (2013.01); H04L 43/028 (2013.01); H04L 43/06 (2013.01); H04L 63/1425 (2013.01)] 26 Claims
OG exemplary drawing
 
1. A system for managing content delivery network (CDN) performance comprising:
a plurality of points of presence (P.O.P.), wherein individual P.O.P.s provide content responsive to client device requests and wherein individual P.O.P.s generate log information corresponding to collected CDN performance metrics associated with providing the content response to the client requests;
one or more computing devices maintaining a plurality of data stores, wherein the plurality of data stores maintain information related to collected CDN performance metric information, historical CDN performance metric information and CDN configuration information;
a CDN management component, executed on a computing device having process resources, the CDN management component operable to:
obtain the log information from the plurality of P.O.P.s;
obtain the historical CDN performance metrics information;
obtain the CDN configuration information;
apply the obtained log information, historical CDN performance metrics information, and CDN configuration information as inputs to a plurality of machine learning algorithms to generate a set of target recommendations, wherein individual machine learning algorithms are trained according to an identified anomaly in the operation of a CDN and wherein the generated set of recommendations includes at least one CDN performance recommendation and at least one CDN analysis recommendation,
wherein the CDN performance recommendations include at least one of cache configuration recommendations, compression configuration recommendations, or time to live value recommendations, and wherein the CDN analysis recommendations include information for characterizing a client impact, the information including at least one of origin source errors, performance degradations, or cache performances;
process the set of target recommendations; and
transmit the processed set of target recommendations.