Machine learning based content delivery

A system comprising:

• a data store configured to at least store computer-executable instructions;
• and
• a hardware processor in communication with the data store, the hardware processor configured to execute the computer-executable instructions to at least:
• determine that a request for a first network resource corresponds to a first group of content requests associated with a first content serving computing device based on a grouping criterion, wherein a first content request in the first group of content requests includes a request for a second network resource different than the first network resource;
• route the request to the first content serving computing device;
• obtain a metric that measures cache performance associated with the request, wherein the metric comprises a weighted average of a cache hit rate of a cache component of the first content serving computing device and a latency of content delivery in response to the request;
• update the grouping criterion using at least the obtained metric that measures cache performance;
• update the first group of content requests using the updated grouping criterion;
• and
• re-associate the first content serving computing device with the updated first group of content requests.

A non-transitory computer-readable storage medium storing computer executable instructions that when executed by a processor perform operations comprising:

• determining that a request for a first network resource corresponds to a first group of content requests associated with a first content serving computing device based on a criterion, wherein a first content request in the first group of content requests includes a request for a second network resource different than the first network resource;
• routing the request to the first content serving computing device;
• obtaining a metric that measures cache performance associated with the request, wherein the metric comprises a weighted average of a cache hit rate of a cache component of the first content serving computing device and a latency of content delivery in response to the request;
• updating the criterion using at least the obtained metric that measures cache performance;
• updating the first group of content requests using the updated criterion;
• and
• re-associating the first content serving computing device with the updated first group of content requests.