More

    Oracle Employs AI to Reduce Analytics Costs

    Oracle today announced it has made available a tool dubbed MySQL Autopilot that employs machine learning algorithms to automate the management of its MySQL HeatWave service, an in-memory query acceleration engine for the MySQL Database Service running on the Oracle Cloud Infrastructure (OCI). 

    The company also unfurled MySQL Scale-out Data Management, which can improve the performance of reloading data into HeatWave by up to 100 times by expanding cluster sizes to 64 nodes, which Oracle claims enables organizations to process up to 32TB of data per node.

    Oracle is also now sharing benchmarking code that it claims show the MySQL HeatWave service provides 13 times better price/performance than Amazon Redshift configured with AQUA and 35 times better price/performance than Snowflake. That capability is significant because it doesn’t require organizations to replace an open source relational database that is already widely employed to run transaction processing applications. The MySQL HeatWave service is designed to process analytics alongside those transactions.

    MySQL Autopilot applies many of the AI technologies Oracle has developed for its namesake database to the open source database that Oracle gained when it acquired Sun Microsystems more than a decade ago. It automates provisioning, data loading, query execution, and failure resolution in addition to sampling data and collecting statistics on data and queries that optimize memory usage, network load, and execution times employing models constructed using OracleML, a framework the company created to automate the building of AI models.

    MySQL Autopilot makes MySQL Heatwave the only cloud service that leverages AI to optimize queries in a way that reduces the total cost of analytics, says Nipun Agarwal, vice president of research and advanced development for Oracle. “You don’t need to move the data,” he adds.

    With each successive query, MySQL Autopilot learns more about how to optimize MySQL HeatWave to the benefit of all the users of the service, added Agarwal. It also predicts the number of HeatWave nodes required for running a workload by sampling data, which Agarwal noted eliminates the need to manually guess how to size a cluster.

    Other benefits include the ability to optimize the load time and memory usage by predicting the optimal degree of parallelism for each table being loaded into HeatWave and predict the column on which tables should be partitioned in-memory to optimize performance.

    Finally, MySQL Autopilot can determine the optimal representation of columns being loaded into HeatWave, estimate the execution time of a query prior to executing it, optimize propagation of changes to the data management layer; prioritize short running queries over longer ones, and auto recover nodes and reloads of data if necessary.

    Also read: AI and Observability Platforms to Alter DevOps Economics

    Retaining Customer Base

    Heatwave is at the core of an Oracle effort to thwart rivals that are targeting customers that previously standardized on Oracle data warehouses. Heatwave, for example, is part of a portfolio of data lake services that Oracle now provides on top of its cloud platform.

    It’s not clear to what degree MySQL Autopilot will increase the overall appeal of HeatWave. However, it’s apparent Oracle is making a concerted effort to surface a variety of cloud services at price points that make it difficult for existing customers to migrate to another platform. Those efforts should also put significant pressure on AWS, Snowflake, Microsoft, and Google among others to reduce costs in response by, in part, applying AI more broadly.

    In the meantime, the amount of data heading in the cloud continues to exponentially increase. The challenge IT teams now face is how to process and analyze all that data without breaking the proverbial bank. Many cloud services may appear inexpensive at first blush, but when that monthly bill comes due many organizations are finding themselves spending more money than ever to analyze data albeit at a better price/performance ratio.

    Read next: Designing Valuable Data Governance Frameworks

    Mike Vizard
    Michael Vizard is a seasoned IT journalist, with nearly 30 years of experience writing and editing about enterprise IT issues. He is a contributor to publications including Programmableweb, IT Business Edge, CIOinsight and UBM Tech. He formerly was editorial director for Ziff-Davis Enterprise, where he launched the company’s custom content division, and has also served as editor in chief for CRN and InfoWorld. He also has held editorial positions at PC Week, Computerworld and Digital Review.

    Latest Articles