New benchmarks demonstrate power of ScyllaDB’s close-to-the-hardware NoSQL ‘scale-up and scale-out’ database
Palo Alto, Calif. — Nov. 5, 2019 — ScyllaDB has done it again. ScyllaDB, the real-time big data database that stunned the industry with performance of millions of operations per second on a single node, has reached an impressive new milestone, the company announced today at its 2019 ScyllaDB Summit user conference.
Independent tests show a cluster of ScyllaDB servers reading 1 billion rows per second (RPS)—performance previously far beyond the capabilities of a database using persistent storage.
“Everyone thought you’d need an in-memory database to hit numbers like that,“ said Glauber Costa, VP of Field Engineering at ScyllaDB. “It’s a testament to the power of ScyllaDB’s close-to-the-metal architecture. No other database comes anywhere close to this kind of performance from storage. For users, it means dramatically faster applications and analytics, and lower costs for hardware and maintenance.”
Bare-metal platform provider Packet partnered with ScyllaDB to conduct the test on 83 of its powerful n2.xlarge servers, each boasting 28 physical cores. The benchmark populated the database with randomly generated temperature readings from 1 million simulated temperature sensors that reported every minute for 365 days, producing a total of 526 billion total data points. Scanning 3 months of data, some served from cache and some from storage, ScyllaDB reached speeds of 1.2 billion data points per second. Scanning a full year’s data, all from persistent storage (no caching), ScyllaDB scanned the entire database at an average rate of 969 million data points per second. With 1 day of data, scanning everything from memory, ScyllaDB achieved 1.5 billion data points per second.
“These results demonstrate how ScyllaDB is able to help users get the most out of their infrastructure, no matter the scale,” said Jacob Smith, co-founder and CMO at Packet. “After distributing load across thousands of cores and dozens of machines, ScyllaDB didn’t lose a step — it kept performing in near real-time, even as more servers were added.”
Unlike competing NoSQL databases, ScyllaDB leverages the full power of modern hardware — such as Packet’s bare-metal servers — with its unique shared-nothing, lock-free and shard-per-core architecture, which allows it to scale up with additional cores, RAM and I/O devices. By contrast, Cassandra’s ability to scale up is limited by its reliance on the Java Virtual Machine, which keeps Cassandra from interacting directly with server hardware.
The performance ScyllaDB demonstrated in these benchmarks has implications for real-world applications. For example, analytics jobs that previously took all day to complete can now be run continuously to provide intraday reports. Organizations can analyze their data with Apache Spark, Apache Storm or Dask using large range scans or even full table scans to discover vital emergent trends or historical patterns.
For more information on ScyllaDB or to try ScyllaDB for yourself, please visit scylladb.com.
ScyllaDB is the real-time big data database. Fully compatible with Apache Cassandra and Amazon DynamoDB, ScyllaDB embraces a shared-nothing approach that increases throughput and storage capacity as much as 10X that of Cassandra. Comcast, Starbucks, Ola Cabs, Samsung, IBM, Grab, MediaMath, AppNexus, Investing.com and many more leading companies have adopted ScyllaDB to realize order-of-magnitude performance improvements and reduce hardware costs. ScyllaDB was founded by the team responsible for the KVM hypervisor and is backed by Bessemer Venture Partners, Eight Roads Ventures, Innovation Endeavors, Wing Venture Capital, Qualcomm Ventures, TLV Partners, Magma Venture Partners, Western Digital Capital and Samsung Ventures. For more information: ScyllaDB.com.Media Contact:
415 848 9175