Building a Unified Entitlement Data Model for License Fulfilment

Abstract

The license rights are stored in a number of systems by software vendors and a customer and purchase were described differently by each system. This loophole postpones the issuance of licenses and led to access control errors. The paper is a single entitlement data model which is a union of customers, contracts, products, entitlements, licenses and activations on a single canonical structure. The model employs entity resolution and blocking to resolve records of customers in four source systems. The design is tested to test a simulated experiment, using 48000 customer records and 72400 entitlement records. A nearest neighbour graph blocking coupled with supervised pruning achieves 99.37 percent of the candidate comparisons, and 95.2 percent of true matches. The corresponding stage is matched with the F1 score 0.941. Entitlement completeness increment to 0.96 and average fulfilment time dropped to 9.8 hrs. The findings demonstrate that a joint model with regulated matching enhances speed and accuracy of license fulfilment.