Shubham Sawarkar
Intel teaming with Snowflake to provide customers with performance and increased flexibility across multiple public clouds with Snowflake’s Data Cloud using Intel Xeon Scalable processors and Intel Xeon-based cloud instances.
Snowflake’s Data Cloud is built on top of multiple cloud infrastructures, allowing for essentially infinite size and concurrency. The high-performance Intel instances serve as a common denominator across cloud service provider availability zones across multiple public clouds, allowing Snowflake to design a flexible cloud solution for its customers.
“Collaborating together, Intel and Snowflake provide optimal performance to our joint customers. For us, that’s the goal. Whatever we can do to improve our price-performance will benefit our end customers, and it’s great to collaborate with Intel to accomplish just that,” said Justin Fitzhugh, vice president of Cloud Engineering at Snowflake.
Intel and Snowflake are starting with performance testing across clouds and hardware instances to better match up the best performing Intel-based instances for certain Snowflake workloads, as part of their drive to innovate new cloud platforms. Snowflake will be able to optimize instances with multiple cloud providers in the near future thanks to insights. This helps Intel create and produce chips with processor features that are beneficial to specific workloads over time.
Intel’s team is collaborating to look at all of Snowflake’s product features that help joint customers. Intel teams are improving Snowflake’s search engine optimization when used to search semi-structured data like JSON as part of testing and proof-of-concept activities. Intel has also improved the balance between data ingestion and query performance; for example, certain query times have been lowered from 7 minutes to 10 seconds.
The value of unlocking Intel architecture value with the latest products and platform innovations through software optimization and ecosystem enablement, according to Arijit Bandyopadhyay, chief technology officer of Intel’s Datacenter and AI Group, is valuable and benefits Intel’s mutual enterprise customers.
“When those we collaborate with grow, Intel grows,” Bandyopadhyay said. “Intel’s engineering teams help disruptor innovators exploit the latest Intel products together with leading software optimizations and silicon engineering to optimize critical workloads across deployment models. Together with teams like Snowflake engineering, we can reach new heights in SQL query acceleration while innovating on the next-generation hyperscaler instance designs across major cloud providers.”
Ramesh Arakere, director of Intel’s Enterprise Software Ecosystem Engineering team, said: “Intel has a long history of innovation and deep collaboration with leading software companies. Continuing with that tradition, the addition of Snowflake to the Intel Disruptor Program will help accelerate the pace of innovation in the data analytics ecosystem.”
Intel aspires to be the silicon platform of choice for applications operating from edge to cloud by collaborating widely with independent software providers to improve software everywhere. Intel is also enhancing and enabling open-source technologies used in Snowflake’s software stack, such as Foundation DB, to accelerate workloads targeted for Intel’s CPUs, providing customer use cases with performance advantages.
