News

Amazon Redshift Rules in Periscope Interactive Analytics Test

• By David Ramel
• 10/21/2016

Data analysis and visualization company Periscope Data tested leading cloud-based data warehouse offerings in order to better advise its customers on which technology to pick and gave the nod to Amazon Redshift over Snowflake and BigQuery.

"After investigating Redshift, Snowflake and [Google] BigQuery, we found that Redshift is the best choice for real-time query speeds on our customers' typical data volumes," the company said in an undated and unbylined blog post titled "Interactive Analytics: Redshift vs Snowflake vs BigQuery," apparently published yesterday.

Periscope provides a data analysis platform that uses SQL queries to hook into databases and data warehouses to produce analytic reports, specializing in data visualization.

In attempting to estimate the respective performance of the three offerings in real-world use cases, the company tested with queries written by Periscope analysts on 1 billion rows of internal page visit logs, weighing in at 150GB of uncompressed data. Periscope said its customers typically work with tens-of-millions to tens-of-billions of rows of data.

The company said it conducted the cost/performance data warehouse research with a focus on real-time, interactive analytics in order to better guide its customers who ask for help in picking a cloud platform to work with Periscope's offerings.

"Our customers want to know where to get the fastest query times, how much data their data warehouse can handle, how much it will cost them. We want to make sure we provide the best recommendations to our users when they ask for our help navigating the data warehouse space, so we put Redshift, Snowflake and BigQuery to the test," the company said.

Going beyond just raw performance data, Periscope conducted initial run-time tests to measure fastest possible speeds and then factored in costs to better serve its budget-conscious customers.

"Bolstered that the warehouse solutions were producing comparable performance, we stepped into a wider set of queries," the company said. "Speed is important, but the deciding factor for many users is their budget. The real question we want to answer is how much performance will be available at different price levels."

Much of the company's report focused on a comparison of Snowflake and Redshift. Periscope actually uses Redshift in its own platform and recently partnered with Snowflake to support its warehouse solution on the Periscope Data Platform. It also tested against BigQuery, which it said is used by "a vocal minority" of Periscope customers.

In addition to performance and cost, the company examined technical models and pricing models of the three offerings.

"For technical teams managing workloads of ranging from the millions to the few billions, Redshift is our winner on both price and performance," the company concluded in a "takeaway." Furthermore, "The pricing model is easy to understand, prevents unexpected fluctuations in cost, and query speeds are available at lower price points for daily-active clusters."

Periscope also published an analysis of the data warehouse management capabilities of the three offerings, though it didn't pick a "winner" in that report.

"BigQuery, Snowflake and Redshift all have Web-based consoles where you control your data, clusters, user management, query logging and system analytics," the company said. "In this post we will visit the management console of each system and the supporting tools for managing your warehouse deployments."

Go to the Data Warehouse Management: Redshift, BigQuery, and Snowflake post for more on that analysis.