US 12,229,191 B2
Systems and methods for in-situ graph-based data processing
Steven Schilders, Columbus, IN (US)
Assigned to Infosys Limited, Bangalore (IN); and InvertIT Inc., Columbus, IN (US)
Filed by Infosys Limited, Bangalore (IN)
Filed on Mar. 30, 2023, as Appl. No. 18/128,984.
Application 18/128,984 is a continuation of application No. 18/121,149, filed on Mar. 14, 2023.
Claims priority of provisional application 63/442,682, filed on Feb. 1, 2023.
Prior Publication US 2024/0256606 A1, Aug. 1, 2024
Int. Cl. G06F 16/24 (2019.01); G06F 16/25 (2019.01); G06F 16/901 (2019.01); G06T 11/20 (2006.01)
CPC G06F 16/9024 (2019.01) [G06F 16/252 (2019.01); G06T 11/206 (2013.01); G06T 2200/24 (2013.01); G06T 2210/36 (2013.01)] 20 Claims
OG exemplary drawing
 
19. A method for loading executable graph-based models, the method comprising:
obtaining, by processing circuitry and based on a first identifier, a first manifest state and a second manifest state, wherein the first manifest state and the second manifest state both comprise the first identifier;
generating, by the processing circuitry, a first manifest and a second manifest from the first manifest state and the second manifest state respectively, wherein the first manifest comprises a first node state storage location, and the second manifest comprises a second node state storage location and an overlay identifier;
obtaining, by the processing circuitry, an overlay manifest based on the overlay identifier, wherein the overlay manifest comprises the overlay identifier and an overlay node state storage location, wherein each of the first node state storage location, the second node state storage location, and the overlay node state storage location corresponds to a respective physical storage medium, and wherein a first node state, a second node state, and an overlay node state are persistently stored at the first node state storage location, the second node state storage location, and the overlay node state storage location, respectively, at the respective physical storage medium;
obtaining, by the processing circuitry, the first node state from the first node state storage location, the second node state from the second node state storage location, and the overlay node state from the overlay node state storage location;
generating, by the processing circuitry, a first node based on the first manifest and the first node state;
generating, by the processing circuitry, an overlay node based on the overlay manifest and the overlay node state;
generating, by the processing circuitry, an executable node based on the second manifest and the second node state;
associating, by the processing circuitry, the executable node with the first node and the overlay node; and
generating, by the processing circuitry, an executable graph-based model comprising the executable node.