Magnit
Sticker
Magnit logo

Opening a new sales channel and improving the existing infrastructure

Magnit is the Russian retail market leader in geographical spread and the number of shopping units. It has more than 21,000 stores in 65 regions. In 2020, the company continued building omnichannel customer experience by launching express delivery.

Tasks
and requirements

Integration with the express delivery services follows a standard schema. The retailer sends a catalog of goods, including descriptions and the number of available stores, to the delivery service. The app then displays the catalog as an online storefront to make sure that the app users always have access to the current product range.

However, the infrastructure and business processes at Magnit were originally designed for offline business needs. The catalog of goods was stored in Magnit’s main database, and information on stock balance was stored locally in every store. For this reason, it was impossible to implement the usual integration schema: Magnit’s digital environment had no integration points for that.

Information on the stock balance in a specific store is transferred to the main database every few hours: during stocktaking and debit and credit accounting. Such a speed is more than enough for delivery planning and meeting customer demands of offline retailers. However, express delivery services update the online storefront in real time, which requires a much higher speed of information exchange.

The retailer needed a software interface that would serve as the middle layer between the company’s IT systems and user applications. Amid the changing conditions of the grocery retail market, one of the most important factors for Magnit was the speed of launching express delivery (time-to-market).

  • Task 1
    2 months to develop the solution
  • Task 2
    Merging data from the main DB and the stores
  • Task 3
    Implementing the solution in the current infrastructure
  • Task 4
    Making the solution quickly scalable
  • Task 5
    Integrating with a number of express delivery services
  • The solution should retrieve the data from the central repository and the stores, merge the data and transfer it to user applications in real time.
  • Tight launch schedule: no more than a couple of months should pass from the moment business requirements are specified until the system is released into production.
  • The project must be incorporated into the existing infrastructure.
  • Ensure solution scalability: Magnit planned to launch express delivery from retail stores in several dozens of cities across Russia by the end of 2020.
  • Magnit was also planning an integration with a number of delivery services: Delivery Club, Yandex.Eats, and the company’s own express delivery app.

Solution

The new software interface is based on the Tarantool in-memory computing platform, which became the middle layer between Magnit’s information systems and the partners’ external applications. The project is based on Tarantool Data Grid. In this Tarantool implementation, deployment issues are initially solved out-of-the-box, which significantly shortens the project timeline.

Due to the tight timeline, the Tarantool team dealt with many business process issues on the fly. For example, the team learned what data already exists in the systems, what information is located in which system, how often that data should be retrieved, and so on. As the project progressed, it became clear that trading via online services is impossible without quantization, and the corresponding data was absent in Magnit’s information environment. Quantum is the minimum lot for different goods. The rules of quantization and other terms of trade for specific products were set up in the Tarantool system. The product catalog was developed and optimized for user applications in 30 days.

Solutions diagram

The solution is deployed on the VK Cloud Solutions platform, thus eliminating the need to purchase equipment.    

As a result, Magnit obtained a solution based on Tarantool Data Grid. This solution collects and organizes the catalog with goods, prices, and stock balances that it gathers from the company’s information systems. It transmits the information to the delivery services in real time. The number of stores connected to the delivery services increased tenfold in the first month—from 20 stores at the beginning of September to 200 in October. The solution allows connecting new applications to the system without any revision or modification of the apps: once a service is connected, it starts reading the necessary information from Tarantool.

Results

  • 20k
    RPS

    Operational solution workload is 20,000 RPS

  • 2
    months

    Research and development

  • 30
    days

    Time to implement access to the Tarantool-based product catalog

  • 400
    stores

    Good scaling: in the first month of operation, the number of connected stores increased from 20 to 200, and in the second month, up to 400

Still have questions?
Leave us a message, and we will get back to you!

Other
success cases

How we combined all data for the investment business of Alfa-Bank
Alfa Bank Logo
Sticker

Project of the Year 2020

The project won the 2020 GlobalCIO Project of the Year award in the category "Infrastructure creation and modernization"

Get
a consultation

Order
a demo

Thank you for your request

Tarantool experts will contact you shortly

Thanks
Thanks