Key achievements in the strategic programs in 2017
The purpose of the UFS program is to create a unified service standard in all channels in which we interact with clients: in bank offices, in the mobile app, the browser and by telephone. The UFS is based on the cross-channel principle, involving the ability to identify the Sberbank client in all channels. The program is designed to help employees work more efficiently and to make the customer’s experience more pleasant and convenient.
In 2017 we developed a new version of the platform – UFS 7.0, with greater reliability and productivity by supporting deployment mode in a multiblock architecture and Stand-In mode, ensuring higher fault tolerance and seamless upgrades of the functional subsystems of UFS.
We have begun mass implementation of a new targeted development process of the UFS using the Development Tools. This will allow a single team to implement a solution for all channels, by reusing as much as possible existing objects and services, reducing the number of errors by auto-generating standard functional units, and reducing the time required to train new employees.
In 2017 we received a certificate from Rospatent of state registration of the Unified Front-End System platform, confirming our intellectual property rights and securing additional legal protection for software.
|Application for consumer loan||Sberbank Online (web)|| |
|Opening and adding to brokerage accounts||Sberbank Online (web)|| |
|Remote reserving of corporate accounts||Bank website|| |
|Payments and transfers for premium segments||Sberbank First zones in offices|| |
|Mobile work place for CIB client managers||Sberbank Online (mobile app)|| |
|Mobile work place for Retail Business unit employees||Sberbank Online (mobile app)|| |
| Test operation in offices: ||Offices|| |
|Sale of insurance and retail products by outgoing calls (Telemarketing)||Contact center|| |
| Test operation in the Internet bank channel: ||Sberbank Online (web)|| |
The DevOps practices implemented for all UFS applications. The use of DevOps reduces the amount of time needed to update applications and makes it possible to avoid errors caused by manual setting of parameters. Up to 80% of code was covered by automatic unit tests, which led to the number of errors at the testing stage falling by half.
In 2017 The Bank determined the quality standard for the UFS platform. Its performance is monitored based on 12 metrics. The use of this standard has helped to cut the number of errors at the testing stage in half, and allowed us to eliminate 50% of errors within 8 hours.
For the professional growth of UFS employees, 850 have been trained in a special program. A non-material incentive program UFS-bonus is in place for active colleagues. Contests are held as part of the working process, allowing us to accelerate the implementation of a new product through the promotion of healthy competition.
The Bank strives to attract talented employees into the UFS program. To this end, in 2017 we:
- started a mentorship program for UFS developers and analysts;
- opened the UFS Design School, the first graduates of which are already interning in The Bank;
- organized the “Advancement” IT conference and technological competition to find innovative ideas and concepts for mobile client services;
- UFS program specialists took part in internal and external conferences and meetups, aimed at attracting the best specialists;
- on the Habrahabr website we maintain a blog, which already has 36 original articles, with more than 310 thousand views.
The program Business Development Support Platform (BDSP) is a universal tool to create business applications. Development of the platform employs the modern technologies of In Memory Data Grid – technologies for distributed automatic computing in the memory of applications that work with large amounts of data in real time.
BDSP is implementing at the Bank the concept of a Marketplace – an online space that brings together significant amounts of different information that clients need.
|Improvement of internal system processes|| |
|Business Hub|| |
|Product Factories|| |
The Data Factory program plays a significant role in the transformation of The Bank into a data-driven company. The foundation of the Data Factory is the creation and development of data services and data storage infrastructure. The program is intended to meet the following objectives:
- increasing the quality, completeness and availability of data;
- ensuring the reliability of business-critical services;
- increasing the quality and speed of management decision-making;
- reduction of time costs on the launch and change of products and services;
- receipt of additional profits from the use of data analysis;
- reduction in the cost of receiving and possessing data;
- monetization of data;
- behavioral analysis of clients and employees to adjust the strategy for working with various segments.
In 2017 we put into full-scale commercial operation key systems for working with big data using Hadoop technologies – Data Cloud and Data Laboratory.
Currently, information has been loaded into the Data Cloud from 34 systems containing data from both internal and external sources. In 2017, on the basis of data received in the Cloud the first data marts with information on individuals and corporations were created. Data marts are layers of a data warehouse according to the different areas. Also the first data domains of the Unified Semantic Layer (USL) of the Data Factory were created. By 2020 we plan to ensure the receipt by the Cloud of all needed consumer data, export all data necessary for business in the form of a data marts and complete the construction of the USL.
Experiments and data research are being conducted on the basis of the Data Laboratory, including the creation of the infrastructure to work with data science models that use intelligent data analysis methods. This work supports the Sberbank initiatives in the field of artificial intelligence.
Data Processing Center
The purpose of the Reliability 99.99 program is to increase the availability of client services regarding the Bank’s critical automated systems and supporting infrastructure:
In 2017 a great deal of work was done and substantial results were achieved in this area. We implemented a georeservation of part of the automated systems in order to guarantee the safety of client data and full restoration of the automated systems operations at the backup facility.
This year the key system duplication regime Stand In (99.99%, 24/7) was implemented for all critical automated systems. This regime ensures that client services are maintained in a limited functionality regime agreed on with the business in the event of breakdowns in the automated systems and when performing technical maintenance work. This allowed us to avoid more than 530 hours of system downtime.
We launched in commercial operation the pilot zones for the automated systems. The use of pilot zones minimizes the risks of implementing a new business functionality.
Other important activities of the reporting year:
- All CIB systems in a short time were migrated from the leased data processing center to a more reliable target facility with no service interruption. A tech-refresh of some of the equipment was performed at the same time. The result was a substantial reduction in lease payments and improved reliability of service.
- We also put the automated system for capacity management into commercial operation, which allows us to forecast growth in the load on systems and the need for equipment resources.
- We revised the main processing system of the Way4 system, which allowed to increase the potential peak load to 5100 transactions per second (the current recorded maximum is 3.2 thousand transactions per second).
- The automatic retry of transactions in the event of an unsuccessful operation of adding cash to a card in self-service terminals and the automatic return of funds if no response is received in the payment system for acquiring operations were also put into commercial operation.
- We completed the migration of client payment servicing to a centralized target solution, which allowed to abandon the non-target facility. In addition we reduced the time for the automated processing of ruble payment orders of legal entities to 2.5 minutes.