Aerospike Inc. today unveiled an innovation strategy that leverages the scale and performance of its real-time data platform to enhance and streamline query capabilities on massive data sets. 

The strategy and first series of innovations—revealed during Aerospike Digital Summit—remove the bottlenecks from today's traditional data architectures that struggle to keep up with data ingest at the edge and artificial intelligence (AI) and machine learning (ML) model drift due to managing data across multiple disparate systems. 

Today's “Reimagine Realtime” Summit presentations showcased innovations and integrations in Aerospike Database 5 that enable massive data sets to run instantly at the edge and the core of the enterprise, enabling data models to be quickly updated with fresh data. 

• Set Indexes provide efficient access to a Set within an Aerospike Namespace. This feature allows fast queries of records within a Set in a petabyte scale database. 
• Expressions Enhancements extend read and write operations with expressions, and new functionality through composition, moving processing closer to the data for greater efficiency. 
• Aerospike Connect for Spark now lets Apache Spark 3.0 applications easily combine streaming and historical data for better decisioning, and reduces the processing time for model iterations and improvements from weeks or days to just hours or minutes. 

“AI and ML applications have an insatiable appetite for data, but traditional databases can struggle to process and combine both streaming and system of record data and to deliver the predictable performance needed for the best decisions,” said Srini Srinivasan, chief product officer and founder, Aerospike. “Today’s enhancements represent a continued expansion of the Aerospike database platform to build upon our strength of acting in real time upon billions of transactions and make it even easier to build and deploy applications for real-time inference-based decisions with lower server footprint.” 

Today’s release of Aerospike Connect for Apache Spark 3.0 follows last month’s release of the Dockerised version of Aerospike Connect for Presto. The Aerospike Connect product family integrates Aerospike Database 5 with popular open-source frameworks, including Apache Spark™, Kafka, Presto, Pulsar, and JMS™. These technologies drive modern data pipelines and processing, enabling organisations to use rich data sets from across the enterprise to build and deploy modern data architectures without custom coding. 

“Every organisation competing for data supremacy requires speed and petabyte scale to improve their decision velocity on massive data sets,” noted R Ray "Wang, Constellation Research CEO. “As petabyte data sets emerge as the norm for organisations expanding their AI and ML initiatives, customers expect their data platforms to improve ease of use and deliver greater efficiency.”

Last year, Aerospike released Database 5 with enhanced Cross-Datacenter Replication (XDR), enabling data to be dynamically routed between two or more geographically distributed clusters. It also added global expressions, which easily routes just the right data to the right target at the right time. The dynamic, fine-grain control of expressions optimises server, cloud, and bandwidth resources.