Case: Bank Data Warehouse Solution
Client: Beijing Agriculture Commercial Bank (BACB)
Background:
Because of the large volume of historical data in BACB’ s database, they experienced performance issues with their OLTP application. In order to improve performance and efficiency, the bank decided to separate its legacy historical data from its online transaction data. This would create a highly optimized database for its daily online transactions, improving response time and performance, and a large data warehouse that would house all the historical data.
Challenge:
BACB had a large volume database (3 TB) which continues to grow exponentially each month.The BACB in Beijing currently has 2.5 million user accounts, 700 branch offices and 800K loan accounts, but this number is growing so fast that the bank can’t keep up.
Client required the project to be completed within 5 months -including design, development and implementation.
The large volume of concurrent users in the system, including regular individual users, branches, and administrative users, presented a huge challenge to a seamless transition. In a given day, the system had about 800 active online user accounts while it could expand to 1500 users during peak hours.
Response:
HIME built a team of 6 software engineers - 4 C programmers, 2 Java programmers - 1 data architect, 1 QA and 1 project manager
Purged all historical legacy data from the existing system, and migrated it o a newly built data warehouse system running on an IBM DB2 database server
Built an ETL process to run on a nightly basis to move data from the online database into the data warehouse
Built a highly scalable data warehouse to allow other external systems to use it for data mining applications
HIME partnered with IBM who provided the IBM technologies used in this project:
- IBM DB2 9.00 under AIX operating system
- IBM onDemand (Content Management)
- IBM Datastage 7 (ETL tool)
- IBM Cognos (Reporting Solution)
Outcome:
HIME’s implementation successfully met the requirements for BACB. Because of the success of this implementation in 2006, HIME was awarded PHASE II of this project to optimize the data warehouse and implement multi-dimensional analysis capability in 2007.
Sales Win Factors:
- HIME won this bid based on the following criteria:
- HIME’s strong financial industry experience in designing and implementing data warehouses.
- HIME’s understanding of the banking business application provided a huge advantage over its competitors.This allowed HIME to propose the most complete technical solution while still completing the project in the required timeframe
- Partnering with IBM enabled HIME to adopt the right tools and products to implement the solution, a key success factor
Computer systems architecture:
- 1. CAS Credit Approval System
- 2. TMAS TextMessaging Approval System
- 3. CCBAVS Centralized Client-Bank Account Verification System
- 4. LPS Loan Protection System
Download PDF 
