US 12,033,085 B2
Replica reliability
Vijayasuriya Ravi, Glen Allen, VA (US); Christina Garcia, Midlothian, VA (US); Raja Kavuru, Richmond, VA (US); Sathiyamurthy Thiruvengadathan, Glen Allen, VA (US); Lakshmi Ph Kommaraju, Glen Allen, VA (US); Mallikarjuna Rao Sambaraju, Glen Allen, VA (US); Nathan Robert Oster, Burke, VA (US); and Philip Austin Kedy, Oakton, VA (US)
Assigned to Capital One Services, LLC, McLean, VA (US)
Filed by Capital One Services, LLC, McLean, VA (US)
Filed on Dec. 1, 2021, as Appl. No. 17/539,691.
Prior Publication US 2023/0169355 A1, Jun. 1, 2023
Int. Cl. G06N 5/00 (2023.01); G06N 5/02 (2023.01); G06Q 20/20 (2012.01)
CPC G06N 5/02 (2013.01) [G06Q 20/202 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer implemented method for processing and verifying real-time calculations in a cloud computing environment, the method comprising:
receiving, by one or more computing devices and at a first data center, a request from a point of sale system to perform a real-time calculation, wherein the real-time calculation is used to make automated decisions about a transaction;
transmitting, by the one or more computing devices, the request to a decision handler housed in the first data center;
determining, by the decision handler, which of a plurality of decision engine queues coupled to the decision handler via application programming interfaces (APIs) to further transmit the request to;
transmitting, by the decision handler, the request to a decision engine queue from the plurality of decision engine queues;
determining, by the plurality of decision engine queues, which decision engines to scatter the request to by prioritizing the decision engines with excess compute resources;
scattering, by the decision engine queue, the request to at least two decision engines,
wherein the at least two decision engines perform the real-time calculation in parallel, and
wherein a first decision engine from the at least two decision engines is in the first data center and a second decision engine from the at least two decision engines is in a second data center geographically separate from the first data center;
receiving, by the decision engine queue, a result of the real-time calculation from a decision engine from the at least two decision engines;
transmitting, by the decision engine queue, the result to the decision handler; and
transmitting, by the decision handler, the result to the point of sale system.