Imply 3.3: Optimizing analytics spend and improving time to insight on freshest data

Imply, the real-time intelligence company, announced Imply 3.3, featuring enhancements that help customers optimize their analytics spend and improve time to insight on their freshest data, while extending Imply’s leading real-time analytics performance to a broader set of queries. An Imply 3.3 download or cloud trial is available on the Imply website.

Enterprise leaders in a range of industries use Imply to deliver self-service real-time analytics to their business users, make business intelligence interactive and exploratory, and create data-driven applications for their customers. They’ve wanted the ability to query multiple data sets directly using standard SQL operations while not sacrificing performance, and now they can.

Imply 3.3 takes advantage of new SQL JOIN support in Apache Druid 0.18. The addition of JOIN operations broadens Druid’s performance advantage over data warehouses and data lake query engines by leveraging Druid’s architectural advantages such as advanced indexing and horizontal query distribution.

Druid’s innate query speed advantage over data lake query engines was demonstrated last year by researchers at the University of Minho (Portugal). Druid displayed a 10X to 59X advantage over Presto and was 110X to 190X faster than Apache Hive.

Support for JOIN operations reduces cloud data storage volumes and compute costs, and enables broad adoption of self-service analytics. Previously, multiple data sets would have to be “flattened” into a single table which included redundant data and made updates expensive.

Now multiple data sets can be used “as is,” simplifying data pipelines and creating substantial savings by reducing storage costs, data ingestion costs and maintenance costs.

For one Imply customer, a Global 2000 retailer who uses Imply to optimize procurement and product mix in real-time, JOIN support will reduce storage requirements by half. It will also greatly reduce their compute costs, as updates will require a fraction of the processing previously required.

“Digital transformation has spawned huge amounts of continuously flowing data. Our customers’ challenge is to bring that data to bear on day-to-day decisions, cost-effectively,” said Fangjin Yang, chief executive officer and co-founder of Imply.

“Our latest release greatly improves the cloud computing economics of real-time intelligence, while maintaining best-in-class performance, so that companies can extend analytics beyond the analyst, to business users, while maintaining fiscal responsibility.”

In this release, Imply also added query laning to improve resource utilization and reduce costs. Query laning works like an HOV lane. It provides prioritized access to a subset of resources for urgent queries. Query laning ensures that interactive time-sensitive queries are never blocked by longer-running reporting queries.

Examples where these enhancements lower total cost of ownership (TCO) include:

• Lower storage costs: denormalizing tables with hundreds or thousands of dimensions can create high storage and compute costs.
• Lower data loading costs: datasets whose dimensions change frequently no longer have to be completely re-ingested with each change.
• Lower data engineering costs: existing queries from BI tools such as Tableau or Looker can be used “as is”, without rewriting JOIN queries.
• Lower compute costs: query laning makes more efficient use of computing resources by consolidating mixed workloads onto the same cluster.