SAP Combines In-Memory Engine With Hadoop
A new in-memory query engine designed to boost interactive analytics capabilities on Hadoop has been added to SAP HANA along with other new cloud platform services.
SAP HANA Vora software released this week aims to leverage and extend the Apache Spark execution framework to boost the performance of Hadoop. The query engine is designed to target distributed data to provide contextual awareness while improving “business process awareness” across enterprise applications and analytics, the company said Tuesday (Sept. 1).
Vora aims to extend in-memory computing to distributed data along with online analytical processing (OLAP) for businesses within the Hadoop ecosystem. The software is also designed to expand access to data as a way to combine corporate and Hadoop data. Hence, the company is pitching Vora as a way to “bridge the gap between corporate data and big data.”
The mash-up capability based on the Spark SQL data source API is intended to let users create projections from enterprise data sources to pump up datasets. Spark SQL semantics also can be used for “drill-down” analysis.
Vora is targeted at the financial services, healthcare, manufacturing and telecommunications sectors, SAP said. Among the possible use cases for Vora are identifying and mitigating risk and fraud by spotting irregularities in financial transactions and customer data.
Another is optimizing bandwidth and improving quality of service by analyzing traffic patterns to avoid network bottlenecks. Finally, Vora is promoted as a platform for scheduling preventive maintenance or improving product recalls by combining and analyzing records on materials and services along with supply chain sensor data.
The query engine is also being positioned as a way to conduct OLAP processing directly on large datasets using in-memory technology and stored in Hadoop. Previously, SAP said, mining large datasets for contextual information using Hadoop was a challenge.
Despite its ability to store and access huge data volumes at lower cost, Hadoop “is not as well suited to the fast, drill-down nature of today’s business questions,” SAP noted. Vora leverages data hierarchies that enable OLAP analysis of Hadoop data along with enhancements in Spark SQL to help improve analytics “across all the data in enterprise applications, data warehouses, data lakes and edge sensors,” the company claimed.
Meanwhile, the in-memory query engine running on the Spark execution framework is designed to compile queries to speed up processing across various nodes. That feature is intended to accelerate and simplify OLAP analysis of large datasets. Moreover, SAP said data moves much faster between HANA and Apache Spark.
The company said SAP HANA Vora is scheduled for released to customers in late September. A cloud-based developer edition is scheduled for release at the same time.
Vora was released on conjunction with a new SAP HANA cloud platform aimed at speeding application development.