Otkritie is a too-big-to-fail Russian bank, one of the top 10 largest credit institutions in Russia. It has more than 10 thousand large corporate clients, 510 thousand small and medium-sized businesses, as well as more than 3 million private customers. Otkritie has a comprehensive portfolio of digital products.
The bank's digital system, which includes services for internal and external customers, maintains more than 57 million accounts and processes data in all physical branches of Otkritie Bank. During the merger of several banks under the Otkritie brand, a sharp increase in the volume of transactions required an upgrade of the system's throughput capacity and higher speed of data exchange between the root systems and user services.
To ensure uninterrupted data delivery and improve the user experience, Otkritie developers addressed three challenges:
The developers needed to:
To address these objectives, the bank's team compared offers and products from different vendors. The short list consisted of Redis, Tarantool, and HazelCast. The products were compared based on the following criteria: price, technical support in Russia, sharding, persistence, and performance. Redis was not a good fit due to the high cost of the corporate version, and HazelCast did not have technical support in Russia. Only the Tarantool in-memory computing platform met all the criteria.
It took only two months to implement the project and fully introduce it in all Otkritie Bank branches. We set up a server with Tarantool in two data centers and refined the middle layer services so that they could access data from Tarantool. Conducted load testing and moved on to pilot operation in a production environment.
Tarantool provided the following benefits:
The platform can handle peak loads that are 20 times the average daily volume of requests
The time to update balance information on customers' accounts was reduced to two seconds instead of dozens of seconds
The data access platform was created in two months without changing the company's IT infrastructure
The reading load from the bank's backend system has decreased by 85%
Tarantool experts will contact you shortly