Aerospike, a NoSQL database vendor, has released a set of extensions that enable better data integration and accelerate data analysis for machine learning workloads.
The connector allows users to easily move various data sources in and out. Aerospike database..Update is the evolution of the company First release The March 2019 Aerospike Connect updated connector with the first versions of the Spark and Kafka connectors also benefits from the Aerospike 5.1 database update that was generally available in July.
James Curtis, an analyst at S & P Global, says vendors are using Aerospike Connect to provide users with low-friction entries into their existing environment.
“Aerospike is part of a major trend that enables NoSQL vendors to analyze, but to date most NoSQL databases have focused primarily on operational / transactional workloads,” Curtis said. Stated. “The company’s Spark connector, which includes updates, is a big part of that strategy.”
Aerospike Connect enjoys the benefits of data center replication
The Aerospike 5.0 database debuted in May 2019 and was updated to version 5.1 in July.
Enhanced connectors use the updated features of the latest Aerospike database release to help improve real-time updates and performance for all connectors.
One of the key additions to the Aerospike Database 5 Series is support for cross-data center replication (XDR). In addition to XDR, the Aerospike Database 5 series has introduced globally distributed transactions. Both new features help users run Aerospike on multiple sites and in the cloud for high performance in a variety of applications, such as payments.
Spark improvements help AI models
The Spark connector in the Aerospike Connect update has been improved to accelerate performance. AI model generation..
Srini Srinivasan, Chief Product Officer and Co-Founder of Aerospike, said that if a developer is running an algorithm for AI model generation, all the data analyzed in memory will be used to quickly generate the model. Tends to try to fit.
Prior to the introduction of the new Spark connector, users tended to copy data from the Aerospike real-time database to another database, such as: HBase, Then run the Spark process. The problem with this scenario is that the user needs to copy the data. This requires time and resource calculations.
Srinivasan said he used the new connector to build dataframe-based access from Spark to the Aerospike database. The new connector now coordinates Spark parallelism with Aerospike parallelism. The end result is that the data is not copied to the supplemental system, allowing users to generate AI models faster and with more real-time data.
Apache Pulsar is coming to Aerospike Connect and Presto is coming next
Aerospike used to have Kafka and Spark connectors, but the Pulsar connector is brand new.Pulsar Event streaming A common alternative to Apache Kafka.
“There was a lot of activity in the pulsar space and we received a lot of demands for it,” said Srinivasan. “We feel that pulsar collectors are used in large quantities on a large scale, based on some of the inputs we have seen on the market.”
In addition to Pulsar, Aerospike is now Presto Currently in beta connector. Presto is an increasingly popular open source SQL query engine and is often considered a competitor to Spark.
Aerospike Connect improves integration with NoSQL databases
Source link Aerospike Connect improves integration with NoSQL databases