Case Study: NUVIAD achieves scalable, lower-cost, near-real-time analytics with AWS Glue

Using Amazon Redshift Spectrum, Amazon Athena, and AWS Glue with Node.js in Production

Nuviad, a mobile marketing platform that processes massive ad-transaction streams, faced rapidly growing data (their Amazon Redshift cluster grew from 3 to 65 nodes and CPU hit ~90%) and needed a lower-cost way to store less-frequently queried data while keeping results fresh for near–real-time analytics. To solve this, they extended Amazon Redshift with Redshift Spectrum and built an ETL pipeline using AWS Glue (alongside Amazon Athena, Kinesis Firehose, and AWS Lambda) to move and catalog data on S3 in Parquet format for immediate querying.

By using AWS Glue to run hourly Spark jobs that convert CSV to Parquet and update partitions for Redshift Spectrum/Athena, Nuviad dramatically reduced the data scanned per query (from 135.49 GB for CSV/GZIP to 2.83 GB with Parquet) and cut query times — simple queries averaged ~12.6s on Spectrum+Parquet versus 21.5s on Redshift (~40% faster), while complex queries ran ~43.1s on Spectrum+Parquet versus 220.8s on Redshift (about an 80% improvement). The AWS Glue–enabled workflow automated partitioning and validation, lowered query costs, and preserved near–real-time performance for Nuviad’s users.

Nuviad

Rafi Ton

Founder and Chief Executive Officer