Sprinter

 

Improved performance and cost optimization by migrating to Google Cloud

Sprinter is the meeting point for people who live for sports. For those people who get up in the morning, take the kids to school, work 8 hours, and come home to put on their sneakers and go out for a run. For those heroines who take time out of where there isn’t any to go to the gym, and know that a good pair of sneakers are more flattering than a pair of high heels. For all those who choose comfort when it comes to dressing, without sacrificing their style.
Sprinter, ever since the first store in 1995 in Valencia, has expanded throughout the country surpassing 100 physical stores. Nowadays, it has more than 2600 employees, including stores, central services and logistics, all with an ambitious expansion plan for the coming years.
In its Ecommerce B2C we find sports and casual fashion, accessories and sports equipment for multiple sports, such as running, fitness, cycling, football, paddle or tennis, etc..
The customer needed a self-scaling re-architecture and optimized performance and costs along with the strategy of consolidating with a unique provider and that is why we opted for the migration of AWS to Google Cloud Platform, greatly improving in costs and service monitoring thanks to their partner Ackstorm.

Solution

Migration to Self-Scaling Design with Docker, CloudFiles Object Storage System

Thanks to the know-how of the Ackstorm team, the system architecture is improved through the use of Google Cloud Platform technology, as well as improved configuration of elements and 24/7 monitoring and management services without affecting the already existing code.
In order to do this, a smaller and more stable infrastructure is created, consisting of a typology of instances defined and tuned to a specific type of task. The addition of computing is done only by scale-out group and increasing the synchronization and speed of access to Redis via socket.
Ephemeral computing and data safeguarding also applies. (Pets vs. Cattle: The Elastic Cloud Story @randyvias).
The database system also features computer data consistent with Master-Slave replication, configuring sizes and proportions of Slave instances different from the originals and thus performing double tasks, serve reading queries and replicate behind Master changes.
To manage the necessary number of slaves in the PostgreeSQL database, the use of an internal balancer is suggested.
Ultimately, provision is added to a DEV or STAGING environment with a minimum element for testing data flows and application through a self-provisioned on-off system.

Want to know more about our case studies?