Главная / Projects / DIVIDING THE DATABASE INTO OPERATIONAL AND ANALYTICAL SECTORS (2015)

DIVIDING THE DATABASE INTO OPERATIONAL AND ANALYTICAL SECTORS (2015)

GOALS AND OBJECTIVES

BUSINESS OBJECTIVE

To optimize and improve the reliability of the automated banking system (ABS).
IT OBJECTIVE

To separate and migrate operational and analytical data to different storage sites.

SOLUTION

A Parallel Migration Scenarios
Oracle Certified Master

IMPLEMENTATION

ABS downtime of no more than three hours could be tolerated, therefore the separation and migration procedure was thoroughly developed and tested several times. During the course of the project, changes were constantly made to application software, and the version of the database itself was radically changed requiring several adjustments to separation and migration mechanisms.

Data separation and migration were performed strictly during dedicated timeslots for technical work. Thanks to this approach, as well as the joint efforts of the bank's developers and Jet Infosystems experts, including Oracle Certified Master, the array reorganization did not affect user work.

Each stage included a plan for rollback to the original state, and migration mechanisms accounted for the relationships between shared databases: small-size tables were migrated in a single stream, while parallel migration scenarios were used for tables of several terabytes. After database division, the complex continues to function as an integral whole, serving all incoming requests just as it did previously.

PROJECT RESULTS

The division of the database, which was originally more than 25 TB, makes it possible to optimize and significantly increase the reliability of the ABS operation. Operational and business intelligence data are now stored and maintained independently, taking account of differing load profiles and availability requirements.

In addition, the backup time has been significantly reduced, the efficiency of hardware use has increased, data recovery in the event of a failure is now almost eight times faster, and it has become possible to optimize the costs of data storage and systems backup.

Replication of the ABS to the backup site now requires lower bandwidth channels, and analytical data is stored in less expensive, Midrange arrays. This has freed up Hi-End disk space which can therefore be used for further expansion of the bank’s systems.

25 TB
Size of the entire database
550 branches
Includes the bank network
8 times
Faster is now the data recovery in case of failures
3.3 million
Private clients are served by the bank

CUSTOMER REVIEW

Operational data made up only one-eighth of our total data volume, yet due to their presence we had to serve the entire database as highly critical. For a long while now, a separation solution has been obvious, but there was also a significant "but": While operating in a single space, data had grown together to such an extent that the necessary parsing carried a considerable risk. So as to resolve this problem, we needed a reliable partner to whom we could entrust our business-critical systems; Jet Infosystems was just the company for the job, as we knew given that we had already implemented several complex, joint projects.

Denis Sotin

Information Technology Director, Rosbank