Since its foundation in 1989, Kuveyt Türk has made significant contributions to the development of participation banking in Turkey with its dynamic structure, corporate governance approach, and innovative products offered to its customers and its expansion abroad. As the only financial institution with two R&D centers in Turkey, Kuveyt Türk has focused on technology and innovation for over ten years. Kuveyt Turk has developed numerous products, including the BOA Banking Platform, XTM, Senin Bankan, CebimPOS, Online Finance System, PowerFactor, and API Platform. Kuveyt Turk exports the technologies developed by its subsidiary Architect to 20 countries worldwide. Kuveyt Türk consolidated its 10th place in the overall banking industry in Turkey and its top ranking among participation finance institutions in terms of asset size. As part of its successful digital transformation journey, Kuveyt Türk focuses on open banking and fintech partnerships. The Bank has data-sharing activities with app developers thanks to over 90 APIs provided on the API Market Platform.
There are approximately 1,000 DLL changes in about 50 channels and 100 different applications in deployments of Kuveyt Türk’s main banking software. Code changes are made in thousands of files between the deployments carried out in one-month periods. Hundreds of requests and calls are entered between the two deployments in the demand call system. It is a huge problem to monitor code quality, security, and performance changes between these versions. Sometimes, steps to improve code quality and security can reduce the system’s overall performance. At the same time, the size and complexity of the system can increase.
The SonarQube application, which Kuveyt Türk uses in code quality, produces outputs for code analysis, code review, code quality, and code coverage, but it does not help Kuveyt Türk in delta analysis. Likewise, it does not provide personnel-based reports by analyzing the commit histories made in the projects. There is no correlation between the code quality and the performance of the work done. Based on this association, it does not make a design proposal for future structures. System performance monitoring and analysis of the change between versions are done with traditional methods. In SmartDelta, Kuveyt Türk applies Performance metrics of cross-version requests for analysis.
Kuveyt Türk Main Banking System (BOA) is a huge framework code and is constantly being developed by more than a hundred software developers. Changes between versions are very important in such large systems. It is very important that these version changes feed back the system development, code development flow, code quality, and system performance. Therefore, our strategic goal in SmartDelta is to correlate the code quality, code change, system performance, and project employee metrics. Also, code quality and performance data between two deployments must be analyzed and outputs must provide feedback on system quality and performance.Kuveyt Türk team in SmartDelta
So far, in SmartDelta, all functional and non-functional requirements have been identified and work packages associated with our requirements have been determined. In terms of being a user scenario provider, the content of the data to be shared with other stakeholders has been clarified. Regarding the functional requirements, the issue was examined in terms of information security, and the method to be followed in data exchange with the stakeholder companies in the consortium was evaluated. The sample dataset was submitted to the IT governance department in order to share with our technology provider in Turkey. For performance data, log and journal data in the main banking system, code repository for code quality data, code scan results, and data on the project and error management system were extracted and analyzed.
The SonarQube application, which we use in code quality, produces outputs for code analysis, code review, code quality, and code coverage, but it does not help us in delta analysis. Likewise, it does not provide personnel-based reports by analyzing the commit histories made in the projects. There is no correlation between the code quality and the performance of the work done. Based on this association, it does not make a design proposal for future structures. In addition, system performance monitoring and analysis of the change between versions are done with traditional methods. Performance metrics of cross-version requests need to be analyzed.
Kuveyt Türk on the Web: