Built from the ground up for maximum performance
The Ocient Hyperscale Data Warehouse™ leverages a unique data warehouse architecture to minimize its footprint on disk and cloud instances while delivering incredible performance for hyperscale workloads.
Ocient’s innovative design across clustering, partitioning, compute adjacent storage, compression, and more delivers the performance at scale required to power next-generation data analytics.
A columnar-oriented storage architecture allows Ocient to read the minimal amount of data off disk possible. By also partitioning data segments by a user-selected compound clustering key and sorting data by time from a selected time column, Ocient produces a primary clustering that is both optimized for a time dimension and for queries based upon the main clustering key. This is called TimeKey clustering, and it is key to Ocient’s blazing fast performance.
By locating NVMe drives adjacent to the processor in a Compute Adjacent Storage Architecture, Ocient is optimized to unleash hundreds of millions of parallel tasks as soon as data comes off the drive. With 12 drives per chassis, data is scanned at over 1.2 terabits per second, or 25 million 4KB random reads per second. Techniques like kernel bypass allow Ocient to minimize bottlenecks and deliver tremendous response time and concurrency so users can realize more value from their data, faster.
We built Zero Copy Reliability, a method of ensuring reliability without replication, into the storage, query, and data access layers of the Ocient Hyperscale Data Warehouse, thereby reducing the storage footprint by up to 80%. We further reduce storage needs by using efficient string compression, run-length encoding, dictionary encoding, and delta encoding to achieve a compression ratio of up to 10X on uncompressed data.
Two complementary datastores are at the heart of the Ocient Hyperscale Data Warehouse, and they’re both always available for query. That’s why users can query data within seconds of loading and keep high volumes of historical data for analysis over longer lookback periods.
Ocient loads data and makes it available for query within seconds via a write-optimized immediate datastore. During loading, Ocient collects enough data for long-term segments that are fully compressed, indexed, and read-optimized. Once this process is complete, the data is automatically moved to long-term storage and removed from the immediate store.
Ocient’s main storage for customer data is a highly efficient columnar read-optimized store with run-length encoding, data encoding, dictionary encoding, and string compression. Data is stored in independent segments, and primary clustering, sort, and secondary indices are stored adjacent to the data. The data is always available for querying, even as it moves between stores.
Join George Kondiles, distinguished architect and co-founder, as he discusses the innovations that lower the footprint of our system and deliver data reliability without replication.
“When I first saw what Ocient could do, the number of systems I wouldn’t have to build just blew my mind. We got faster performance, lower cost, and had less to maintain.”
“I was really impressed with Ocient’s capabilities, particularly in data ingestion and partition tolerance both in terms of loss of a node but more so the recovery of lost data. Ocient has built an impressive product backed by a very impressive team.”
“Ocient has a six-year lead time over any competitor and technology that spans the whole SQL database stack, right down to server memory allocation. This specialization is a formidable technology moat that any competitor would need to cross. Unless the market for trillion row, near-realtime SQL databases heads toward the 1,000-plus customer level then it looks like Ocient is operating effectively on its own.”
“We selected Ocient because of its promising solutions to rapidly and cost-effectively accelerate data-driven innovation at hyperscale, which is critical for enterprises to generate real-time transformational value.”
Learn how you can deploy the fastest data analytics solution out there.