In the first quarter of 2022, Hulu reported 45.3 million paid subscribers, up from 39.4 million in the same quarter of the previous fiscal year. And as more viewers continue to stream, the amount of data being pulled in from the platform exponentially increases.
Hulu processes roughly two billion rows of data an hour, pulling in metrics from all of their user platforms including browsers, mobile apps and streaming devices like Roku and Firestick. This high data ingest (and high volume of users) means Hulu is a data-intensive application. These mass amounts of data gathered allow the streaming giant to glean insights on their performance, and provide the best customer experience possible.
Hulu’s original data architecture used Druid and Storm for data aggregation and analytics. One of the biggest Druid installations in the world, Hulu used it to analyze traffic, and predict regional issues — and to combat slow loading and buffering that often leads to bad streaming experiences.
Hulu continued to struggle with the Druid-Storm system and eventually, experienced an outage during the 2018 Super Bowl — an event that drew more than 100 million viewers. Not only did Hulu’s system go down, but the company had no idea what was going on. Even more, they couldn’t gather the data and analyze it fast enough.
Re-Powering a Data-Intensive Application With SingleStore
Following the outage, Hulu needed to ensure a catastrophic event like that never occurred again. Looking for an alternate technology, Hulu needed a system that would meet their visibility requirements, be more stable and lower total cost of ownership (TCO). The answer? SingleStore.
With SingleStore powering their data architecture, Hulu was able to streamline the design of its data ingestion. Before using SingleStore, Hulu used a six-step process for ingestion — a full bag of tricks to complete the ETL to get data in. The Druid-Storm system was complicated, and led to several performance bottlenecks that required workarounds. Now, Hulu streamlines processing with the SingleStore Pipelines feature, a single database command that automatically brings in data to dramatically reduce platform complexity. Not to mention when Hulu moved to SingleStore, they discovered they could cut hardware costs in half — needing only a quarter of the memory and much lower CPU resources, as well.
With SingleStore, Hulu created a real-time, ongoing stream of insights to predict and forecast the quality of their service. Instead of reacting without visibility, Hulu has a real-time view into what’s happening across its streaming service. Today, Hulu processes roughly a hundred terabytes of performance data per day with a dramatically simplified pipeline — and far less technology to manage.
SingleStore is the cloud-native database built with speed and scale to power your data-intensive applications. Book your demo today.