Browse Prior Art Database

Effective Matching of record in Federated Environment

IP.com Disclosure Number: IPCOM000228528D
Publication Date: 2013-Jun-17
Document File: 3 page(s) / 40K

Publishing Venue

The IP.com Prior Art Database

Abstract

The dislcosure proposes a technique for identifying relevant records and matching them effectively in a federated environment consisting of plurality of hubs communicating with each other. The participating hubs might be implemented on a variety of architectures with various restrictions which might impede in returning matching data of high confidence. This disclosure proposes techniques for formulating a complete search request which will in turn return data with high quality from all the other participating hubs. The technique also includes setting configurable levels of thresholds based on requirement and being able to specify the hub to be used for matching the individual data. Also, a technique is proposed to join similar data returned from the individual participating hubs which are returned as output.

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

Page 01 of 3

Effective Matching of record in Federated Environment

Networked computer systems operating in a federated environment may allow participants to exchange data and derive value from the data which other systems might possess concerning an entity or a subject of common interest. However, since the participating systems might be implemented on a variety of architectures including but not limited to, the search/ match algorithm deployed, the attributes defined and the weights assigned to them and the confidence threshold, the data returned from the participants depends to a considerable extent on the search criteria. Also, there can be restrictions on the searchable data and the data returned due to

policies and restrictions to certain participants which might result in return of false positives whilst false negatives not being returned. But, the caller of a particular hub can make use of the system which issues the search to effectively formulate the search request to ensure other

participating systems return consistent data with better quality.

Prior Art

Patent/Publications Main Idea

US 2008/0244008 A1 Probabilistic Federated Match is a capability of finding a unique record across all the hubs / data nodes with highest confidence starting from the hub that issued the federated match request. Currently there are no systems that are available which can deliver such result as the systems exists today has some kind of identifiers to find the records across the hubs.

The disclosure tables a way of getting a unique record from the local and ensuring matching data with highest confidence is returned form the other participating hubs. In a federated environment any hub can issue a match request as the primary hub and obtain data from the other hubs if configured properly. The configuration consists of the details of the other hubs participating in the federation. Also, since the hubs can be on a variety of architectures, the configuration can include details on data mapping between the current and remote hub, configuring various thresholds for the hubs and also specifying the hub to be selected for confidence matching. Since, each hub can have its own configuration different from the others and not a common configuration, implicitly the original caller and the primary hub that issues further requests to remote hubs sets the levels of data quality and quantity returned. The other main advantage comes from the fact that since data is not stored anywhere either in the primary or the remote hubs or any third party centralized hub and is returned on-the-fly, security policies existing on various hubs would still hold and the data returned would also be current data with no overhead in maintenance of data coming from remote hubs.

In order to get a unique record from all the hubs with high confidence, the first important requirement is for the search cri...