American Apparel Company

Executive Summary

Carter's Inc, an American designer, and marketer of children's apparel wanted to design single point access for data that was available from multiple sources & stored in multiple formats. Their goal was to enable forecasting and projections for sales, inventory, and pricing. They partnered with ITTStar to create a retail datastore, a combination of Data Lake in AWS S3 and a Datawarehouse in AWS Redshift which helped streamline all data & facilitated the implementation of forecast models.

Customer Challenge

The customer was using traditional on-premise environments which had limitations to handle their complex workloads especially since their data was scattered. The data was stored in multiple systems and multiple formats. Executing any forecasting models against the source systems was also impacting the performance of other applications and processes. Revamping this environment with new hardware, software licenses, and other integrations would also have a huge cost impact. Hence the customer wanted to move to a serverless setup using cloud services & create a data lake which would enable making all the data accessible for implementing forecasting models that create projections for sales, inventory, and pricing from one single location.

Partner Solution

On understanding the customer requirement, ITTStar implemented the S3 Data Lake and Redshift Datawarehouse solution by creating the AWS Glue ETL jobs to extract data from all the identified sources. Data transformation rules were applied where required. S3 events and CloudWatch events were used to trigger the Step Functions that orchestrated the execution of multiple Glue jobs based on dependencies. SNS topics, alerts & subscriptions were created for alerting the business users on the successful completion of job failures. This architected solution was configuration-driven to allow business users to add more source objects to be captured in Data Lake and Redshift. The focus was on security, performance, and cost throughout the implementation of the solution. In addition to ETL jobs, forecast models were implemented for pricing, sales, and inventory using the objects fetched into their retail data store. Custom reports were created to provide record counts comparisons between source and target and to identify any schema changes implemented on the source systems.

Why AWS?

AWS provides a host of services that enable architecting cost-effective and elastic solutions. AWS Serverless ETL services like Glue provides scalability, S3 provides the lowest cost option for the Data Lake, Redshift provides users a single place to query data, thus allowing users to join data stored in different locations for further processing without actually having to physically load all the data into Redshift or having to unload all the data into S3. Both Athena and Redshift queries support standard SQL query language which most IT and business users are familiar with. Services like IAM, KMS, CloudWatch, and SNS span the entire solution providing desired security, monitoring, and alerting. All in all, AWS was the best fit solution.

Why customer chose the partner?

Database creation, Data warehousing, and Data Science related technologies are a strength of ITTStar which are well supported by technical staff that is certified in various AWS Cloud services. The customer was lured by ITTStar’s experience in architecting and delivering a solution by keeping a focus on data security, performance and cost, making them a good consulting and implementation partner for all infrastructure and data related engagements.

Impact & Results

Delivered a secured, configuration-driven and performance-oriented solution Using serverless services helped save costs & thus convert them from CapEx to OpEx Implemented monitoring and exception handling to minimize manual intervention. 24X7 monitoring of the ETL pipelines enabling quicker resolution of issues if any Implementation of forecasting models to improve overall performance