Browse Prior Art Database

Adaptive Batching Algorithm for File Transfer and B2B System Based on Predictive Analysis

IP.com Disclosure Number: IPCOM000238975D
Publication Date: 2014-Sep-29
Document File: 2 page(s) / 41K

Publishing Venue

The IP.com Prior Art Database

Abstract

An adaptive batching algorithm for file transfer and business-to-business (B2B) system based on predictive analysis is disclosed.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

Adaptive Batching Algorithm for File Transfer and B 2

Disclosed is an adaptive batching algorithm for file transfer and business-to-business (B2B) system based on predictive analysis.

Most B2B file transfer systems employ protocol adapters whose function is to write files to a file repository. Once the files are completed written to the repository, a schedule job runs periodically to collect the files and process them. In general these schedulers are triggered by a timer, meaning that when the time pops, the system picks up batches of documents to work on. The problem is that these time based systems are sub-optimal. it is often the case that files either sit around for a longer than they need to before being processed, or that systems are busy constantly checking for new files when no new files exist.

A typical B2B file transfer system has many components including a processing engine and file transfer engine. Such systems can and often do keep track of historical data about when files come in, how large files are, etc. This information allows them to be able to predict both arrival times and completion times. These system also have insights into in-flight file transfers, specifically they know how large the incoming file is, the transfer rate and how much data has already been received. Based on this information, and based on historical data, the system can calculate an 'expected' time that the file transfer will complete.

Two improvements to the existing batch processing methodology are disclosed. First, the batch processing algorithm is augmented so that before it completes its processing, it determines if any in-flight file transfers are predicted to end wi...