US 11,782,763 B2
Resolution of tickets in a multi-tenant environment
Shubhojyoti Ganguly, Howrah (IN); Dipangshu Mukherjee, Hyderabad (IN); Koushik Chakraborty, Kolkata (IN); Raghunandan Bhat, Mangalore (IN); Sampath Kumar Sunkesala, Bangalore (IN); and Pranay Kumar, Bangalore (IN)
Assigned to Oracle International Corporation, Redwood Shores, CA (US)
Filed by Oracle International Corporation, Redwood Shores, CA (US)
Filed on May 7, 2019, as Appl. No. 16/404,798.
Prior Publication US 2020/0356412 A1, Nov. 12, 2020
Int. Cl. G06F 9/50 (2006.01)
CPC G06F 9/5027 (2013.01) 20 Claims
OG exemplary drawing
 
1. A digital processing system comprising:
a random access memory (RAM) to store instructions; and
one or more processors to retrieve and execute said instructions, wherein execution of said instructions causes said digital processing system to perform the actions of:
creating a machine learning model designed to predict gross jobs based on historical information in tickets and gross jobs executed for resolution of said tickets, wherein said machine learning model is a decision tree comprising a plurality of nodes, each node being associated with a corresponding keyword, and wherein each of node of only some of said plurality of nodes are associated with a corresponding gross job;
receiving a ticket containing description of issues for a first tenant from a ticketing system, wherein said description contains keywords;
determining a gross job of a plurality of gross jobs based on said keywords of said description, said gross job representing a class of jobs, wherein each job of said class of jobs is suitable for resolution of said ticket in a corresponding tenant configured to be served by a respective combination of computing resources, wherein said determining comprises:
traversing said decision tree based on the probability of occurrence of said keywords to identify a set of nodes of said plurality of nodes, wherein each node of said set of nodes is associated with a corresponding gross job;
computing, associated with said traversing, a respective confidence value for each node of said set of nodes, the confidence value reflecting a degree of certainty of said corresponding gross job to address the issues in said ticket given the extracted keywords of said description; and
selecting, based on said respective confidence values, a first node of said set of nodes, wherein the corresponding gross job associated with said first node is determined as said gross job;
identifying data specifying a first combination of computing resources currently configured to serve said first tenant;
selecting a first target job from said class of jobs represented by said gross job, based on both of said gross job and said first combination of computing resources currently configured to serve said tenant; and
executing said first target job to cause resolution of said ticket for said first tenant,
wherein said determining is performed without using said data specifying computing resources currently configured to serve said first tenant, and wherein said selecting is performed after said determining.