The AtScale Story

AtScale was born from the compatibility and performance issues vendors have with enterprise data lakes. Dave Mariani, CEO and co-founder of AtScale, saw this challenge first hand at Yahoo! where he oversaw the data pipelines for all analytic users. Mariani struggled to connect thousands of business intelligence (BI) users using different BI tools to a Hadoop data lake with adequate performance.

The Solution. Fast-forward to 2016, and AtScale’s ‘universal semantic layer’ takes the place of an OLAP interface and enables business users to connect to any BI tool and to any enterprise data lake. AtScale can reduce operating costs and increase agility and governance. AtScale executives believe that centralized BI platforms are on their way out, and AtScale’s open, flexible strategy is the future.

AtScale Version 5.5

Building on its early success, AtScale recently announced version 5.5 of its Intelligence Platform. Included in this release are multi-level metric support, high availability deployment, and prediction-defined aggregates.

Multi-level metric support. Multi-level metrics enable users to query multiple fact tables with different levels of granularity. For example, one fact table stores data by region and month, while another stores it by store and day. With multi-level metrics, a user can issue a query for regional sales by store and receive an answer that is not a null set.

High availability (HA) deployment. Some organizations use AtScale to power the front-end of their data analytics applications. Before version 5.5, if AtScale went down, these organizations would need to follow directions in a handbook to get their applications back online. With HA deployment, AtScale is deployed on multiple servers, so if a server goes down, the applications continue to run.

Prediction-defined aggregates. AtScale applies algorithms and statistics based on historical user activity to automatically generate aggregates on new data sets. Once new data is available and the prediction-defined aggregates function is on, AtScale generates aggregates before users begin querying the data. For example, a clothing retail chain may create aggregates, such as size, style, product line, etc. Pre-building aggregates using statistics increases query performance and enhances the user experience.

Looking forward, AtScale expects to offer integration with Google BigQuery and Amazon Redshift by the end of the third quarter. Integration with Druid, an open source data store often used for OLAP queries, is also expected by the end of the year.

 To find out more about how AtScale works, click here.