Breaking Data Silos by Building a Centralized Data Warehouse
KOLON FnC, like its name ‘Fashion & Culture’, is a company that creates culture through fashion and aims to bring more happiness to customers’ daily lives. Starting with ‘KOLON SPORT’ in 1973, KOLON FnC has grown into a representative fashion brand house, encompassing not only men’s and women’s, golf and accessories, but also beauty and lifestyle.
KOLON FnC had complex raw data sources from on-premises and another cloud platform, and wanted to break this data silo. Cloocus suggested simplifying the data pipeline and achieved operational and cost efficiency by building a centralized data warehouse based on Google Cloud BigQuery.
Cloocus has carried out the project largely under three main objectives.
- Integrate Data on Cloud environment
Configuring a Cost-Effective Pipeline with Dataproc Workflow
- Construct cost-effective and operation-effective Data warehouse
Configuring an architecture based on BigQuery, a serverless and fully managed data warehouse
- Discover insight easily
Excavating insight easily by utilizaing connected sheets, external tables etc.
Cloocus proposed a data analysis architecture based on BigQuery for integrated analysis of scattered data. to build an integrated analysis environment, and orchestrated the entire data pipeline using Cloud Composer. Data was collected using Dataproc, and the entire data pipeline was efficiently orchestrated with Cloud Composer, a managed airflow service.
In particular, with shared Virtual Private Clouds only available in GCP, Cloocus configured KOLON FnC’s development and operations organizations to centrally control network resources (Subnets, Firewalls, Routing, etc.) and apply consistent policies to environmental targets separated by multiple projects.
- Cloud Composer
- Google Cloud Storage
- Cloud SQL
- Google Compute Engine
- Cloud VPN
- Cloud NAT
- VPC Service Controls
- Access Context Manager
KOLON FnC not only can quickly process various raw data by building an integrated data pipeline in the GCP environment, but also achieves operational efficiency through a fully managed service-based architecture. In addition, Dataproc Workflow Templates significantly reduce analysis costs by automatically creating and deleting Dataproc Cluster when needed.
As a result, analysis costs have been reduced by more than 30% compared to previous architectures, and overall pipeline management and operations are possible with only 10% of the workforce compared to the previous one.
Based on the data loaded in BigQuery, KOLON FnC expects to introduce AI/ML systems such as demand forecasting and supply chain management using BQML and Vertex AI in the future.