Case Study: Large Mobile Telecommunications Company achieves 18x faster ETL and reporting with SQream DB

Saving hours on reporting with SQream DB

Large Mobile Telecommunications Company, a Tier‑1 mobile operator in Asia with roughly 50M subscribers, faced slow, unreliable analytics: tens of billions of daily customer events, a 5‑hour ETL window and 2–4 hour reporting delays on an 80‑node Greenplum MPP cluster that prevented near‑real‑time targeting and ad‑hoc analysis. They engaged SQream to deploy SQream DB, a GPU‑powered analytics database, to support location‑based ad targeting, customer segmentation and faster BI access.

SQream implemented SQream DB on a repurposed 2U server with a single NVIDIA K80 (handling ~40TB) and a 2TB/hr ingest pipeline, removing pre‑aggregation and simplifying ETL. The deployment cut ETL from 5 hours to 20 minutes and reporting from 2–4 hours to about 10 minutes, sped time‑to‑market for reports by ~18×, enabled queries across fresh and historical data at scale, and avoided expanding the Greenplum cluster—saving the operator hundreds of thousands of dollars annually while broadening access to analytics.