MUMBAI, India, Aug. 24 -- Intellectual Property India has published a patent application (202641098919 A) filed by Vellore Institute Of Technology on August 15, 2026, for A Machine-Learning-Driven Risk-Aware Adaptive Retry Control System For Financial Transactions.
Inventors include P Balakrishnan; A Anny Leema; S. Harshitha Bai; Akshyaa Nagarathinam; and Bhavadharini Shankar.
The application for the patent was published on August 21, 2026, under issue no. 34/2026.
Abstract: I/We Claim: 1. A pattern-aware retry optimization system for financial microservices comprising: • at least one financial microservice configured to process monetary transactions in a distributed computing environment; • a pattern-aware retry agent configured to monitor transaction failures and control retry execution; • a pattern learning engine configured to generate a retry success probability based on learned failure patterns; • a transaction risk analyzer configured to compute a transaction risk score based on transaction attributes; and • an adaptive retry policy engine configured to perform retry gating by permitting retry execution only when the retry success probability exceeds a predefined threshold and the transaction risk score remains below a safety threshold, wherein the adaptive retry policy engine includes a stability control mechanism configured to suppress repeated retry attempts during persistent failure conditions, thereby reducing duplicate transaction processing and retry storm propagation. 2. The system of claim 1, wherein the pattern learning engine is configured to classify detected failures into transient failures, intermittent failures, and systemic failures based on observed runtime failure sequences. 3. The system of claim 1, wherein the transaction risk analyzer is configured to suppress retry execution for transactions associated with elevated financial risk or non-idempotent execution states to prevent duplicate monetary processing. 4. The system of claim 1, wherein the adaptive retry policy engine dynamically adjusts retry timing, retry delay intervals, and retry attempt limits based on the predicted retry success probability score. 5. The system of claim 1, further comprising an evidence logging component configured to record retry decisions, execution outcomes, predicted retry success probability scores, and computed transaction risk scores for auditability and regulatory compliance purposes. [0047] A pattern-aware retry optimization framework for financial microservices is disclosed. The framework integrates a Pattern-Aware Retry Agent (PARA) within each microservice to control transaction retry behavior. A Failure Pattern Monitor collects failure-related information, which a Pattern Learning Engine analyzes using machine learning to generate a retry success probability and classify failure types. Concurrently, a Transaction Risk Analyzer computes a transaction risk score based on attributes like value and idempotency. An Adaptive Retry Policy Engine dynamically determines an appropriate retry strategy by gating execution based on the predicted retry success probability and the computed transaction risk score. This system includes a stability control mechanism to suppress repeated retry attempts during persistent failures, preventing retry storms and duplicate processing. An Evidence Logger records decisions for auditability and continuous improvement, enhancing reliability and consistency in distributed financial environments. [FIG. 1]
Disclaimer: Curated by HT Syndication.
