Woori Bank is the flagship unit of Woori Financial, the Republic of Korea’s largest financial group. The bank’s credit card division wanted to shorten the time it took to load online transaction processing (OLTP) data from the accounts system—which processes and stores real-time credit card transactions—into an analysis database. Woori’s staff uses the analysis database to do customer credit rating assessments, run management reports, and generate the intelligence needed to develop effective marketing and promotional campaigns.
Due to rapid growth in the credit card business, the bank needed a stable, scalable, and highly available database platform that could process large volumes of transactional data very quickly. The ability to access credit card data on the day the transaction occurred—rather than a day later—would help the bank develop effective business strategies, such as creating targeted promotions to increase chances of product take-up. The bank also wanted to minimize the cost, time, and effort of transferring existing databases to a new data processing platform, and it needed to ensure efficient storage of up to 1 TB of new data a month.
After implementing Oracle Exadata Database Machine and Oracle GoldenGate on Oracle’s Sun Fire X4170 M2 and Sun Fire X4270 M2 servers, the bank can now load OLTP data in real time, compared to one day in the past. As a result, customer credit-rating assessment tasks can be completed in 15 minutes, instead of three hours; daily marketing target selections can be made in less than 30 minutes, instead of 3.5 hours; and targeted data for business management reports can be generated in less than 10 minutes, instead of five hours. The faster OLTP loading time has significantly increased work productivity of business users and IT administrators.
Faster reporting means 120 power users in the marketing and other divisions can go beyond standardized queries to generate personalized queries in the data mart access layer, using the results to create effective and timely promotional campaigns. In the past, it had been difficult for them to revise and incorporate changes to marketing campaigns in real time because they were working with the previous day’s data. Now, users can immediately access and analyze transactions for specific trends in the data mart access layer and adjust campaigns and strategies accordingly. Woori Bank can also send tailored offers to customers. For example, if someone applies for a credit card, the bank can follow up with an offer for a personal loan on the same day.
Woori Bank used the Hybrid Columnar Compression feature in Oracle Exadata to compress analytical and historical data, shrinking 78 TB of data down to 42 TB and reducing data storage costs by 46%. The bank also deployed Oracle’s StorageTek SL500 modular library system to efficiently store corporate data. The Oracle infrastructure can be expanded without extensive configuration or installation, simply by adding new racks to easily support the bank’s future growth.
In December 2010, Woori Bank’s credit card division and Woori FIS (Woori Financial Group’s IT services business) ran a series of benchmark tests of integrated OLTP/OLAP solutions from four vendors. The solutions were assessed on their performance, based on queries and reports that the bank regularly generates.
The results showed Oracle Exadata offered the fastest performance and the best database integration capabilities. The platform was also optimized for loading and analyzing accounts data in real time. In addition, Oracle Exadata simplified data verification. In the past, verifying the integrity of accounts data was a load-intensive process that could take three days to complete. With Oracle Exadata, verification of all account data takes just two hours.
Oracle Exadata’s Hybrid Columnar Compression feature—which can compress data by up to 10x—was another attractive feature, as it could provide efficient storage and support the expected growth in Woori Bank’s credit card business and data volumes.
Finally, Oracle offered an architecture that can process large data volumes, is easy to scale, and provides quick and stable migration, due to the use of identical databases. This would enable service continuity during the migration process and lower deployment costs.
In May 2012, Woori Bank transferred 7.5 TB of credit card transaction data from the accounts system to the Oracle Exadata platform, completing the task in just 1.5 hours. The arrangement time needed to optimize Oracle Exadata was reduced from an expected 22 hours to 7 hours.
The Oracle Exadata platform is performing stably after the bank set the data compression range and optimized application performance.