Virtualization Technology News and Information
NeuroBlade Demonstrates Acceleration of Analytics Workload by up to 100x, using its SQL Processing Unit (SPU) Technology
NeuroBlade is revolutionizing the speed and performance of query processing on petabyte-sized datasets. NeuroBlade is excited to announce the successful demonstration of its NeuroBlade Hardware Enhanced Query System (HEQS) utilizing KIOXIA CM7 Series Enterprise NVMe SSDs. This groundbreaking solution empowers customers to harness the full potential of high-performance SSD throughput, resulting in query performance improvements of up to 100 times.

Despite improvements, general-purpose CPUs struggle to keep up with the growing I/O throughput, which has increased by approximately 17 times in the past 5 years. NeuroBlade's purpose-built SPU architecture overcomes traditional CPU limitations, eliminating system bottlenecks and enabling high-speed query processing regardless of complexity. This delivers unmatched performance, reduces footprint, and improves energy efficiency. With NeuroBlade, organizations can fully leverage their data assets to identify new business opportunities and deliver enhanced services at a lower cost.

The NeuroBlade HEQS is a rack-mounted server that combines NeuroBlade's innovative SQL Processing Unit (SPU) with high-bandwidth PCIe Gen 4.0/5.0 NVMe flash storage from leading companies, like KIOXIA, and their CM7 Series SSDs. By processing data closer to -storage, the HEQS maximizes the performance of NVMe flash drives, minimizing data movement and accelerating overall system performance for data-intensive tasks. This enables customers to achieve pure linear scalability, scaling their performance to match their largest requirements.

"Our collaboration with KIOXIA strengthens our go-to-market strategy and underscores the growing demand for innovative and powerful solutions," said Elad Sity, CEO and co-founder of NeuroBlade. "This breakthrough empowers organizations to keep pace with their exponential data growth, elevate analytics performance, and gain an invaluable competitive advantage. We owe this success to our exceptional engineering team, who have worked closely with companies like KIOXIA to unlock this level of innovation for data analytics."

The NeuroBlade HEQS seamlessly integrates into a customer's environment, accelerating analytics workloads by 10-100 times while reducing TCO by up to 10 times and power usage by up to 85%. It effortlessly connects to various database query engines without requiring changes to existing data, queries, or code. This significantly enhances performance for analytics workloads such as Business Intelligence (BI), Data Warehouses, Data Lakes, ETL, and more, enabling organizations to achieve substantial improvements in performance, cost, and power efficiency.

"We welcome the collaboration with NeuroBlade and their efforts to address the performance gap in data analysis" said Neville Ichhaporia, Sr. Vice President and General Manager of the SSD Business Unit, KIOXIA America, Inc. "Our CM7 Series SSDs are well-suited for data analytics workloads that require high throughput access to data storage. We are delighted to work with NeuroBlade and contribute to the delivery of an innovative solution that accelerates widely used database applications."

The demo includes NeuroBlade's HEQS G100, featuring 32 KIOXIA CM7 Series NVMe drives, 512 GB memory, and 8 NeuroBlade SPUs. By benchmarking standard TCPH tests with a 10TB Scale Factor, the demo highlighted the impressive results achieved by the NeuroBlade accelerated solution compared to a standard server running software only, equipped with 128 CPU cores. The benefits of the NeuroBlade solution included up to 80% utilization of KIOXIA SSD drives I/O throughput, up to 23 times query performance improvement, and up to 21 times improvement in consistent Query Data Scan Rate (QDSR).

Published Thursday, August 03, 2023 12:12 PM by David Marshall
Filed under:
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<August 2023>