Prediction of cacheable queries

Patent metadata

Field	Value
Publication number	US-12353410-B2
Application number	US-202318520359-A
Country	US
Kind code	B2
Filing date	Nov 27, 2023
Priority date	Nov 27, 2023
Publication date	Jul 8, 2025
Grant date	Jul 8, 2025

Abstract

Predicting cacheable queries is provided. For example, a system integrates one or more processors with a cache store to execute one or more microservices. The system receives a first request including an identifier and a set of key-value pairs. The system predicts a set of requests based on the identifier and the set of key-value pairs. The system identifies based on a comparison with a threshold metric, a subset of predicted requests from the set of requests indicative of one or more subsequent requests. The system constructs a set of labels. The system retrieves, via the set of labels and from a data repository remote from the client service, data for the subset of predicted requests. The system transmits a cache value from the cache store that corresponds to the subsequent request.

First claim

What is claimed is:

1. A system, comprising:
one or more processors, coupled with a cache store, the one or more processors configured to:
receive a first request that includes an identifier and a set of key-value pairs;
predict, using a machine learning model, a plurality of requests indicative of one or more subsequent requests based on the identifier and the set of key-value pairs from the first request;
identify, using the machine learning model and based on a comparison with a threshold metric, a subset of predicted requests from the plurality of requests, the subset of predicted requests comprising predicted request identifiers and corresponding predicted key-value pairs;
construct, using the machine learning model and based on the predicted request identifiers and the corresponding predicted key-value pairs, a set of labels comprising classifications for the subset of predicted requests;
retrieve, using the set of labels and from a database, data for the subset of predicted requests;
configure, in the cache store, the set of labels as cache keys and the data retrieved for the subset of predicted requests as cache values; and
transmit, responsive to receipt of a subsequent request that matches one of the subset of predicted requests, a cache value from the cache store that corresponds to the subsequent request.

2. The system of claim 1 , comprising the one or more processors to:
receive a second request;
predict, responsive to receiving the second request, via the machine learning model, a second plurality of requests based on the first request and the second request;
identify, via the machine learning model and based on a second comparison with the threshold metric, a second subset of predicted requests from the second plurality of requests; and
transmit, responsive to receipt of a second subsequent request that matches one of a group of predicted requests comprising the second subset of predicted requests and one or more of the subset of predicted requests that are below a cache store threshold, a second cache value from the cache store that corresponds to the second subsequent request.

3. The system of claim 1 , wherein each of the plurality of requests comprises a metric associated with a latency of the plurality of requests and a frequency of the plurality of requests.

4. The system of claim 1 , wherein identifying the subset of predicted requests comprises the one or more processors to:
determine a metric for each of the plurality of requests by calculating a product of a latency associated with each of the plurality of requests and a frequency of receiving each of the plurality of requests;
rank the plurality of requests by respective metrics; and
select the subset of predicted requests based on the ranking.

5. The system of claim 1 , comprising the one or more processors to identify that the plurality of requests was marked as cacheable, prior to a training of the machine learning model.

6. The system of claim 1 , comprising the one or more processors to remove the cache keys and the cache values from the cache store responsive to an elapse of a threshold period of time.

7. The system of claim 1 , wherein predicting the plurality of requests based on the identifier and the set of key-value pairs comprises the one or more processors to:
generate, via the machine learning model, a hash from each of the set of key-value pairs according to a predefined hash function;
concatenate, via the machine learning model, each hash with a value of the set of key-value pairs to generate input feature encoding; and
identify, via the machine learning model, the plurality of requests by identifying requests associated with the input feature encoding.

8. The system of claim 1 , comprising the one or more processors to:
determine that a metric associated with the subset of predicted requests exceeds a rate threshold;
generate a schedule of retrieval for the subset of predicted requests; and
retrieve, via the set of labels and from the database remote from a client service, the data for the subset of predicted requests according to the schedule.

9. The system of claim 1 , comprising the one or more processors to suggest at least one of the one or more of the subsequent requests by presenting the one or more subsequent requests in an application operating on a client service.

10. The system of claim 1 , comprising the one or more processors to train the machine learning model to capture a temporal pattern of requested profiles associated with a client service based on historical data associated with a plurality of profiles from entities, wherein each profile of the requested profiles is one of the plurality of profiles associated with an entity of the entities.

11. The system of claim 1 , comprising the one or more processors to predict, via the machine learning model, the plurality of requests based on received requests from a client service during a user session, wherein each received request of the received requests comprises an entity identifier of an entity and a profile associated with the client service, each entity associated with a plurality of profiles comprising a respective profile of each received request.

12. A method, comprising:
receiving, by one or more processors, coupled with a cache store, a first request that includes an identifier and a set of key-value pairs;
predicting, by the one or more processors using a machine learning model, a plurality of requests indicative of one or more subsequent requests based on the identifier and the set of key-value pairs from the first request;
identifying, by the one or more processors using the machine learning model and based on a comparison with a threshold metric, a subset of predicted requests from the plurality of predicted requests, the subset of predicted requests comprising a predicted request identifiers and corresponding predicted key-value pairs;
constructing, by the one or more processors using the machine learning model and based on the predicted request identifiers and the corresponding predicted key-value pairs, a set of labels comprising classifications for the subset of predicted requests;
retrieving, by the one or more processors using the set of labels and from a database, data for the subset of predicted requests;
configuring, by the one or more processors, in the cache store, the set of labels as cache keys and the data retrieved for the subset of predicted requests as cache values; and
transmitting, by the one or more processors, responsive to receipt of a subsequent request that matches one of the subset of predicted requests, a cache value from the cache store that corresponds to the subsequent request.

13. The method of claim 12 , comprising:
receiving, by the one or more processors, a second client request;
predicting, by the one or more processors, responsive to receiving the second request, a second plurality of requests based on the first request and the second request;
identifying, by the one or more processors using the machine learning model and based on a second comparison with the threshold metric, a second subset of predicted requests from the second plurality of requests; and
transmitting, by the one or more processors, responsive to receipt of a second subsequent request that matches one of a group comprising the second subset of predicted requests and one or more of the subset of predicted requests that are below a cache store threshold, a second cache value from the cache store that corresponds to the second subsequent request.

14. The method of claim 12 , wherein identifying the subset of predicted requests co

Assignees

• Adp Inc

Inventors

• Roy Kuntal
• Liao Xinpeng
• Jain Ruchi Jinendra

Classifications

• G06N20/00

  Machine learning · CPC title

• G06F16/3349

  Reuse of stored results of previous queries · CPC title

• G06F16/2453 Primary

  Query optimisation · CPC title

• G06F16/24552 Primary

  Database cache management · CPC title