TL;DR:
- Ocient introduces OcientGeo for geospatial data analytics and integrates ML tools into its Hyperscale Data Warehouse.
- The company focuses on hyperscale workloads without relying on GPUs, emphasizing extreme parallelization.
- OcientML allows machine learning on massive datasets with superior price performance and simplified operations.
- OcientGeo enables geospatial queries on massive datasets directly within the platform.
- Ocient’s commitment to improving price and performance will reshape the data analytics and machine learning market.
Main AI News:
In the realm of data, there’s big data, and then there’s truly massive data, characterized by trillions of rows of information. Chicago-based Ocient has firmly planted itself in this space with its cutting-edge hyperscale data warehouse technology. Today, Ocient unveils a set of groundbreaking capabilities that elevate the hyperscale data platform to new heights, catering to geospatial data analytics, machine learning (ML), and artificial intelligence (AI) requirements.
Embedded within Ocient’s Hyperscale Data Warehouse product, the newly introduced OcientGeo feature boasts an extensive library of geospatial functions and a globally optimized spatial index. With OcientGeo, enterprises gain the ability to seamlessly ingest and process massive volumes of historical and real-time geospatial data, unlocking actionable insights. Integrated ML tools further empower businesses to expedite geospatial AI initiatives.
What sets Ocient apart is its commitment to delivering hyperscale capabilities without the reliance on GPUs. According to Ocient CEO Chris Gladwin, “Our focus is hyperscale workloads, and I would say the average number of elements that are looked at in an average Ocient query, whether it’s SQL, machine learning, or geospatial, is on the order of probably an average of a trillion things.”
Hyperscale Data Analytics: A Flow-Centric Approach
While many organizations turn to GPUs for enhanced performance in accelerated computing scenarios, Ocient takes a distinct approach to fortify its data warehouse. Gladwin emphasizes that the key to their success lies in extreme parallelization. He elaborates, “It’s not at all unusual that at every layer in the stack, there’s a million parallel tasks in flight or more.”
To achieve this massive parallelization, Gladwin underscores the importance of “flow.” He clarifies that in machine learning algorithms for clustering, regression, and classification, the computational operations within a CPU are not the bottleneck. Instead, the bottleneck often revolves around compute density, demanding more computational power per terabyte of data.
Addressing this challenge, Ocient has developed groundbreaking technology to optimize memory and leverage fast solid-state drive (SSD-based data storage systems. As Gladwin puts it, “Our engineers would love to work on GPUs; they’re super cool, but we just haven’t found a need.”
Hyperscale Machine Learning with OcientML
Ocient’s data warehouse originally focused on SQL data queries, and this architectural foundation now extends to OcientML and OcientGeo capabilities. Gladwin emphasizes that the advantages of hyperscale performance, real-time analytics, and data loading that Ocient offers for SQL workloads are equally accessible for ML.
OcientML empowers customers to perform machine learning on datasets comprising billions, hundreds of billions, or even trillions of data points, all while delivering superior price performance compared to alternatives. It incorporates features like workload management to ensure equitable resource allocation across various queries and analyses operating at hyperscale. OcientML seamlessly integrates the ML stack into the Ocient Hyperscale Data Warehouse, eliminating the need for data extraction, transformation, and loading onto a separate platform.
The merits of OcientML extend to enhanced model accuracy through direct interaction with historical and current data, quicker iteration cycles by eliminating data movement steps, and simplified operations by managing SQL and ML within a unified system.
OcientGeo: A Seamless Geospatial Solution
OcientGeo mirrors the pattern established by OcientML, residing at the core of the Ocient Hyperscale Data Warehouse and harnessing the platform’s substantial parallelization capabilities. Gladwin highlights that with OcientGeo, customers can execute geospatial queries, analyses, and functions on massive datasets directly within the Ocient platform, bypassing the need for extensive data extraction. This enables the execution of queries and analyses involving trillions of geospatial data points in a matter of seconds on a massive scale.
Gladwin concludes by saying, “We are still at the beginning of a journey that enables new use cases made possible by dramatically improving the price and performance of hyperscale analytics by tenfold or more.” Ocient’s dedication to redefining hyperscale data warehousing promises to reshape the landscape of data analytics and machine learning for years to come.
Conclusion:
Ocient’s latest advancements in hyperscale data technology, with OcientGeo and OcientML, position the company as a major player in the market. Their focus on extreme parallelization and performance optimization without GPUs sets a new standard. This innovation will undoubtedly disrupt the data analytics and machine learning landscape, offering businesses unprecedented capabilities for handling massive datasets and driving actionable insights.
