I have written before about the continued use of specialist operational and analytic data platforms. Most database products can be used for operational or analytic workloads, and the number of use cases for hybrid data processing is growing. However, a general-purpose database is unlikely to meet the most demanding operational or analytic data platform requirements. Factors including performance, reliability, security and scalability necessitate the use of specialist data platforms. I assert that through 2026, and despite increased demand for hybrid operational and analytic processing, more than three-quarters of data platform use cases will have functional requirements that encourage the use of specialized analytic or operational data platforms. It is for that reason that specialist database providers, including Ocient, continue to emerge with new and innovative approaches targeted at specific data-processing requirements.
Ocient was founded in 2016 to develop an analytic data platform to enable interactive analytics on datasets of hundreds of billions of rows or more by computing up to trillions of rows per second. Data lakes have become a popular approach for storing and processing large volumes of structured and unstructured data but can struggle to deliver the performance required for interactive analytics compared to data warehouses with predefined schema. Ventana Research’s Data Lakes Dynamics Insights research shows that almost three-quarters of organizations (74%) are using both data lakes and data warehouses. Target industries for Ocient include ad tech, financial services, geospatial, government, operational IT, and telecommunications. Existing customers include digital advertising company Basis Technologies, advertising technology provider MediaMath, network-level intelligence provider Gigamon, lawful intelligence specialist SS8, and network performance and security monitoring firm Cubro. Ocient has raised $65 million in venture funding, including a $10 million Series A funding round announced in March 2018. An additional $15 million was added in June 2020, prior to a $40 million Series B round in January 2021. Ocient has used that funding to deliver the Ocient Hyperscale Data Warehouse, which is available as a managed service, either on-premises, in the public cloud, or hosted by Ocient in OcientCloud. In addition to targeting complex analytics on the largest datasets, Ocient is also differentiated by its high-touch approach to customer engagement. As well as providing the Ocient Hyperscale Data Warehouse, the company works with potential customers to design custom schema, test queries and data integration flows, and on the development and testing of the related analytic environment prior to deployment into production, and then manages the software and related infrastructure to ensure uptime and performance. Additionally, the company offers customers the option of what it calls “full solution support,” which sees Ocient taking responsibility for the design and analysis of the infrastructure and queries.
In addition to a commitment to customer service, the high-touch engagement approach is also related to ensuring that customers can configure and manage their software and infrastructure to take full advantage of Ocient’s differentiated capabilities. Although the Ocient Hyperscale Data Warehouse runs on industry standard hardware, it is differentiated from more general-purpose data warehouse offerings by, amongst other things, being optimized to take advantage of compute-adjacent NVMe (nonvolatile memory express) solid state drives, which are designed to deliver high levels of throughput and fast response times, in order to provide the I/O required for interactive analytics on very large datasets. Ocient Hyperscale Data Warehouse provides functionality to enable users to transform, stream or load data directly into the data platform with the ability to transform data as it is ingested to reduce time to query, and it indexes data on ingest with support for multiple types of secondary indexes that can be optimized to match the nature of the data and query. Ocient Hyperscale Data Warehouse is also designed to store and process both structured and semi-structured data, including multidimensional and geospatial data types, and provides Zero Copy Reliability to reduce requirements for disk storage, as well as associated rack space and cooling.
The company recently introduced version 20 of Ocient Hyperscale Data Warehouse, which delivered new capabilities focused on extract load and transform (ELT) data ingestion in which the transformations are performed by the database, rather than loader nodes. Ocient Hyperscale Data Warehouse V20 also delivered hyperscale log analysis via support for large strings and N-gram indexing, as well as the new Ocient Simulator, which is essentially a free version of the database delivered as a Docker image to make it easier for customers to test environments and accelerate development and quality assurance testing. Ocient Simulator enables functional testing but is not meant for production use. Although Ocient Simulator is designed to facilitate customer adoption, it is offered in the context of the company’s high-touch engagement approach, which sees Ocient fully engaged in the four stages of the pilot-to-production process: planning and requirements gathering; architecture selection and deployment; pilot engagement; and production deployment. The Company manages all Ocient Hyperscale Data Warehouse deployments for its customers, whether those are deployed on-premises, in the public cloud (Amazon Web Services or Google Cloud) or hosted in Ocient’s own data center (OcientCloud).
This managed approach, combined with the specific configuration requirements, means that Ocient Hyperscale Data Warehouse will not be suitable for organizations looking for a self-service approach to data warehouse consumption and deployment. Additionally, infrastructure requirements are predetermined and fixed (along with associated pricing), which means Ocient Hyperscale Data Warehouse is targeted at workloads that do not fluctuate and therefore do not require elastic scalability. Nevertheless, given the differentiated functionality and customer engagement approach, I recommend that organizations with the most complex analytics workloads consider Ocient and Ocient Hyperscale Data Warehouse when evaluating their options.