Project Reporting Process Digitalisation
Cloud application to support the digitalisation and automa ...
Client background & business context
Client: MediaMarktSaturn Retail Group (leading consumer electronics retailer)
Project: modernisation of legacy goods management system for improved demand forecasting and replenishment
Challenges
With a distribution network of over 1,000 stores in 14 countries, the MediaMarkSaturn Retail Group needed a more scalable and comprehensive goods management system. After a thorough analysis, we identified the key aspects that needed to be modernised to better respond to the client's business needs. Acting as partners, we collaborated on the implementation of the most viable solution for enhanced cost-effectiveness and future readiness. Some of the challenges faced in this implementation included:
- Finding a suitable digital solution that would help the company have a centralised overview of quantified store data (e.g., stocks, prices, or orders)
- The need for real-time data from stores to be used in other business processes like automatic stock replenishment
- Implementing a scalable future-ready platform: enabling the future development of any needed additional integration and interfacing
- Providing relevant advanced analytics and business insights that are big-data-specific
- Enabling faster data processing by using an in-memory computing engine, especially for the generation of transactions (BEW) data reports and open order calculations
Our approach
For this project, we started by analysing the company’s business challenges and needs. Together with the stakeholders, we decided to proceed with a modernising and modularising approach. This translated into the creation of a central platform capable of ensuring faster access and synchronising real-time data and services across an extended multichannel distribution.
Our specialists took end-to-end responsibility for the architecture, development, testing, rollout, environments, and monitoring of the project.
The new central platform can access both data and services across the company’s multichannel distribution. The implemented in-memory distributed data grid can handle more than 10,000 transactions per second with a latency of around 1-2 milliseconds, thus handling the company’s existing peaks. To scale the interfaces for the outbound channel and the large number of clients, we used two main approaches:
- A pull approach: we used REST microservices that scale horizontally consuming data. This was used especially for synchronising the entire data layer.
- A push approach: we generated asynchronous data streaming with WebSocket that is decoupled from the persistence layer by using ESB (Kafka) topics.
Keeping in mind the future-readiness aspect of the solution, our specialists facilitated shorter deployment cycles and continuous integration.
Technologies
- Java
- Spring
- Angular JS
- Haproxy
- Nginx
- Consul
- Kafka
- Hazelcast
- Postgre SQL
- Swagger
- Python
- Java 10
- Kotlin
Results
The previous ETL-based solution was very slow, taking 5-10 minutes to place an order and often requiring human intervention. The new solution takes approximately 2-3 seconds per request and requires minimal to no human intervention.
Requests to place an order or print a sales document, which were previously done with human intervention, are now backed up by elegant and secured REST API interfaces. This reduces to zero the risk of possible human error.
Also, in the case of new integrations, this approach reduces the risk of system crashes, making it a highly available and scalable system that can be easily extended. The solution enables business partners.
Related case studies
München
Mies-van-der-Rohe-Strasse 6, c/o Design Offices
80807 München, Germany
E-mail: hello (at) accesa.eu
Cluj-Napoca
Constanta 12, Platinia
400158 Cluj-Napoca, Romania
E-mail: hello (at) accesa.eu
Oradea
Cetatii Square 1, Oradea Plaza
410520 Oradea, Romania
E-mail: hello (at) accesa.eu