St. Petersburg Bank develops data technology using Arenadata Hadoop

2024: Enterprise Data Warehouse Functionality Migration to Arenadata Hadoop

On May 6, 2024 Bank Saint Petersburg , he announced that as part of the development IT infrastructures , he had migrated corporate functionality to data stores a product from the Arenadata Hadoop (ADH) Russian manufacturer. As software Arenadata a result of the project, the Bank managed to create and transfer to industrial use a stable, fault-tolerant platform that meets the requirements of Business Critical systems.

The decision to implement the data management platform, along with the launch of a series of projects to centralize disparate storage based on it, has optimized the IT landscape in a number of ways with increased efficiency. It also helped to create scalable solutions for the actively modernized BSPB IT architecture. noted Alexander Rybakov, Senior Vice President of Bank Saint Petersburg

As reported, the Bank has historically had a corporate data warehouse based on Oracle solutions. At that time, it covered the required amount of tasks, but the organization's team understood that the platform reached the maximum of its functionality and ceased to meet the needs of business customers. As a result, it was decided to create an updated data platform based on Apache Hadoop technologies, which would allow you to implement the necessary functionality and horizontal scaling that contributes to the development of the internal infrastructure. The office of CDO Bank "St. Petersburg" chose the product Arenadata Hadoop (ADH) of the Russian developer Arenadata.

{{quote 'author
= commented Yan Guzov, CDO of Bank "St. Petersburg"|An important selection criterion was the guarantee that the product expertise will remain on the Russian market regardless of external conditions. It was understood that Arenadata - a domestic vendor who will provide us with up-to-date versions of his distribution, timely and high-quality technical support in time, and will also be able to train our specialists in various practices in using the product. For May 2024, we see that the solution was strategically correct in the context of today's import substitution tasks.}}

The implementation of the migration project took several years, since the Bank's CDO office and other divisions were tasked with transferring all existing QCD functionality belonging to the Business Critical class to the updated platform.

The possibility of our data platform is that it is built without using the hot data layer provided in the architecture by the BI layer. However, for May 2024, the bank's team is thinking about adding a universal hot data layer, for which it is conducting a pilot project to use the QuickMarts DBMSrenadata (ADQM). noted Gleb Smirnov, owner of Hadoop platform in Data Office

Ensuring data resiliency and availability is one of the top priorities for financial institutions. St. Petersburg Bank has built a platform on the basis of Arenadata Hadoop that meets the requirements for Business Critical systems. believes Stanislav Gabdulgaziev, Arenadata Sales Support Department expert

As of May 2024, many departments of St. Petersburg Bank use the developed data platform for building corporate reporting. With its help, they solve tasks such as managing credit limits on cards, forming pre-approved proposals, and preparing reports on marketing campaigns.

Arenadata Hadoop (ADH) is a full-fledged distribution based on Apache Hadoop, adapted for corporate use. It is designed to store and process semi-structured and unstructured data. Among the key tasks that Arenadata Hadoop solves are:

• Cost-effective storage and efficient processing of data of various formats.
• The ability to scale to petabytes of data using standard equipment.
• Ensure data availability and availability. Data stored on any node is duplicated on other nodes in the cluster. This helps to avoid system shutdown due to hardware and software failures. If something happens to one of the nodes, then there is always a backup copy of the data available in the cluster.