Elastic unveils upgrades for Elasticsearch on Google CPUs

Elastic has announced performance improvements for users of its Elasticsearch platform on Google Cloud, reporting up to 40 per cent higher indexing throughput when running on virtual machines powered by Google's new Axion processor.

The gains were measured on C4A virtual machines, the first VM family using Google Axion – the company's custom Arm-based CPU – and are available to users through Elastic Cloud Serverless, which is now in technical preview on Google Cloud.

Elastic said it used a macro benchmarking framework for Rally with the elastic/logs track to test the maximum indexing throughput on the Axion-powered infrastructure.

"Elastic is driving innovation and cost-efficiency by enabling customers to leverage our Search AI-powered search, observability, and security solutions on Arm-based architecture," said Uri Cohen, vice president of product management at Elastic.

"Google Axion processors augment Elastic's best-in-class capabilities, enabling users to index data more efficiently and improve search performance," he said.

The announcement comes as part of a broader collaboration between Elastic and Google Cloud, aimed at supporting developers, site reliability engineers, and security teams in building and optimising generative AI applications.

Salil Suri, director of product management for Google Compute Engine, said the performance benefits are part of Google Cloud's efforts to offer specialised infrastructure for high-demand workloads.

"We're excited to bring the efficiency and performance benefits of our custom Google Axion processors to Elastic Cloud Serverless," Suri said.

"Powered by our new Titanium SSDs local storage, our first Axion VM family – C4A – provides Elastic users up to 40 per cent higher indexing throughput, underscoring our commitment to continuously deliver workload-optimised infrastructure to power Elasticsearch on Google Cloud."

Elastic Cloud Serverless is designed to deliver more flexible, scalable deployment options for users of Elastic's suite of solutions, which span search, observability and security use cases.

The new performance metrics position Elastic and Google Cloud to better serve organisations seeking faster, more efficient platforms to support data-intensive generative AI initiatives.

Elastic, headquartered in the United States with a presence in Australia, provides real-time data analysis solutions built on its Search AI Platform. According to the company, more than half of the Fortune 500 use its technology for a range of enterprise functions.

The company did not provide a timeline for general availability of Elastic Cloud Serverless on C4A but confirmed that the technical preview is now live for users on Google Cloud.