Browse Prior Art Database

Method and apparatus of conflict reduction in multi-source data synchronization

IP.com Disclosure Number: IPCOM000192673D
Original Publication Date: 2010-Jan-28
Included in the Prior Art Database: 2010-Jan-28
Document File: 7 page(s) / 211K

Publishing Venue

IBM

Abstract

The disclosure includes a method and system to solve the primary key conflictions issue during multi-source database replication by using rule-based approach to dynamically change conflicting key values and propagate them to enable data consistence. In the disclosure, during database replication, firstly tables are linked together by primary key and foreign key relationships to generate a table dependency graph which is a directed graph of tables, then detect the replicate sequence by the table dependency graph. If any conflict is detected, change the primary keys according to predefined combination rule and keep adding the data into the target database. Further if any primary key is changed, all foreign keys of the primary key are changed according to table dependency graph. After the confliction reduction, legacy applications could run without any change.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 49% of the total text.

Page 1 of 7

Method and apparatus of conflict reduction in multi-source data synchronization

Disclosed is a method and system to solve the key conflictions issue during multi-source database replication by using rule-based approach to dynamically change conflicting key values and propagate them to enable data consistence. After the combination, legacy applications could run without any change.

In which a method of database replication with datacombination
Define

relationships

    
Generate table sequence by

relationships
Detect key confliction by combination rules; changethe primary keys by rules; and keep adding the data into the target database If any primary key is changed, all foreign keys of the primary key are changed accordingly
In which a system of database replication with datacombination

Table sequence generator to generate the replicate sequence by

relationships

Table selector to get tables in sequence
Record selector to get next record
Conflict detector/resolver to detect and resolve confliction according to combination rules Recorder manipulator to modify and store records into target database

In database replication field, there are the following replication scenarios:
(1) Insert unique record without conflicts with other data from one data source to the target database.
(2) In multiply data source, there is a record wheredifferent data source includes different fields for the same object and no conflicts among these fields. In this case, create a new data record in target database by merging all data fields.

(3) There are data conflicts during replication.

Multi-source data conflict has the following categories and sub-categories shown in Fig 1. :

1

Page 2 of 7

Figure 1. Multi-source data conflict category

The disclosure focuses on non-key conflict. Prior arts include the following approaches:
(1) "one win" rule approach: in the approach, only one source will win, its data will overwrite other data. All existing database products use the approach

(2) Add new column to differentiate different data source


The following description shows an example of the approach (additional flag approach):
Combine multiple source data to one target table (Fig. 2). The source table all have "ID" column as unique index and will have conflict during consolidation. The solution for this problem is to add a Flag column in both source and target and identify each row with ID and Flag column.

2

[This page contains 1 picture or other non-text object]

Page 3 of 7

Fig. 2 an example for approach 2


(3) Automatic data consolidation (IBM patent US 2006/0036624 A1.

Description of solution (detect the keys from heterogeneous database and combine them into new schema):
- It is a base patent of data replication.
- Combine multiple source data to one target table.The table schema is auto-generated by database access detection
- It detects the relationship of

fr...