Sign up for our daily and weekly newsletters to get the latest updates and exclusive content on cutting-edge AI coverage.
Imagine big data, then multiply it to an extreme where you have trillions of rows. That’s where Chicago-based Ocient thrives with its hyperscale data warehouse technology.
Recently, Ocient introduced new features to its hyperscale data platform, specifically designed for geospatial data analytics, machine learning (ML), and artificial intelligence (AI). The new feature, OcientGeo, is integrated into Ocient’s Hyperscale Data Warehouse product, offering an extensive library of geospatial functions and a globally optimized spatial index. With OcientGeo, companies can process massive amounts of historical and real-time geospatial data to gain actionable insights. Integrated ML tools further accelerate geospatial AI projects.
One of the key promises Ocient makes is to handle these vast data needs using highly optimized storage and processing, without relying on GPUs.
Ocient’s CEO, Chris Gladwin, explained that their focus is on hyperscale workloads. In an average Ocient query, whether it involves SQL, machine learning, or geospatial data, they typically handle around a trillion elements. Unlike many organizations that use GPUs to improve performance for various computing tasks, Ocient takes a different path for its data warehouse. The company achieves this by employing an extreme level of parallelization. At each layer of their stack, millions of parallel tasks are managed simultaneously.
For Ocient, the challenge lies in ensuring enough throughput across the computing stack, including storage and memory. This challenge led to Ocient’s technical edge: optimizing memory and using high-speed solid-state drives (SSD) for data storage. Despite the allure of GPUs, Ocient’s engineers have focused on streamlined CPU operations to maximize output.
Ocient started with SQL data queries, but the same architecture enabling rapid analytics on massive datasets also supports OcientML and OcientGeo capabilities. The benefits of hyperscale performance, real-time analytics, and data loading that Ocient provides for SQL now extend to ML. OcientML allows customers to conduct machine learning on datasets with billions or even trillions of data points, offering a price-performance ratio superior to existing alternatives. It also features workload management to ensure fair resource distribution across various queries and analyses running at hyperscale. By integrating the ML stack directly into the Ocient Hyperscale Data Warehouse, there’s no need to extract, transform, and load data onto a separate platform.
OcientML enhances model accuracy with full interaction with both historical and current data, speeds up iterations by removing data movement steps, and simplifies operations by managing SQL and ML within one system. Similarly, OcientGeo inherits the core advantages of the Ocient Hyperscale Data Warehouse. Users can perform geospatial queries and analyses involving trillions of data points directly within the platform, dramatically reducing query times to mere seconds.
Gladwin emphasized that Ocient is still at the beginning of its journey, enabling new uses that only hyperscale analytics can provide, offering ten times or more improvement in price and performance.