Analyst Perspectives

The publication of Ventana Research’s 2023 Operational Data Platforms Value Index earlier this year highlighted the importance of incorporating analytic processing into operational applications to deliver personalization and recommendations for workers, partners and customers. This importance is being accelerated by interest in generative AI, especially large language models. The emergence of intelligent applications has impacted the requirements for operational data platforms with the need to support real-time analytic functionality. These requirements play to the strengths of vendors of databases that support hybrid data processing such as SingleStore. The company has recently added new features to improve query performance for transactional applications, as well as highlighting support for vector functions that provide similarity search for generative AI applications.

SingleStore was founded in 2011 to create a new relational database management system with an initial focus on in-memory transactional workloads based on its fast data ingest, high-performance memory and indexing functionality. Over time, SingleStore added support for tiered storage by utilizing columnar, disk-based storage and cloud-based object stores. The company’s Universal Storage table type evolved its column store to be capable of performing both transactional and analytic processing.

Interest in hybrid data processing has risen steadily driven by demand for real-time data processing to support the development of intelligent operational applications that deliver personalization and contextually relevant recommendations. Example applications include customer experience, people analytics, finance analytics and marketing and sales effectiveness. Also, fraud detection and prevention in credit approval processes, predictive maintenance and anomaly detection in manufacturing and IoT, and supply chain planning and forecasting in retail and logistics. I assert that through 2026, the development of intelligent applications providing personalized experiences will drive demand for data platforms capable of supporting hybrid operational and analytic processing.

SingleStore has attracted growing interest from customers and investors alike, raising a $146 million F-2 round of financing, announced in the fourth quarter of 2022. The company has used the funding to invest in the development of its data platform, with cloud, analytics, developer and security enhancements added with the launch of version 8.0. The company will also invest in marketing efforts to raise awareness of the use cases for SingleStore, including support for generative AI applications.

SingleStore offers SingleStoreDB, a self-managed distributed SQL database, and SingleStore Cloud, a managed, on-demand cloud database service. Both are designed to support the development and deployment of data-intensive applications, which the company defines by a combination of data volume, query latency, query complexity, data ingest speed and concurrency. In addition to its hybrid data processing functionality, SingleStore also boasts high performance batch and streaming data ingestion with support for programmatic pipelines and multi-format data storage. SingleStore can store data in in relational SQL, JSON and geospatial formats and supports full-text search. Version 8.0 of SingleStoreDB, launched in late 2022, delivered query performance for transactional applications using JSON, real-time and historical monitoring capabilities and support for federated authentication using OAuth. The company also recently announced the launch of SingleStore Kai for MongoDB as part of SingleStoreDB Cloud. SingleStore Kai for MongoDB is an application programming interface for real-time analytics and vector similarity search on JSON data stored in the MongoDB document database, and supports the use of MongoDB-compatible tools, drivers and skills.

The recent excitement around generative AI and large language models has increased attention on vectors and vector search in the data platform sector. Vectors are multi-dimensional mathematical representations of features or attributes of raw data, which could include text, images, audio or video. Vector search utilizes vectors to perform similarity searches by enabling rapid identification and retrieval of similar or related data. Vector search can be used to support natural language processing, as well as recommendation systems that can find and recommend products that are similar in terms of function or style, either visually or based on written descriptions. Vectors and vector search can be used to complement large language models to reduce accuracy and trust concerns through the incorporation of approved enterprise content and data. The recent success of LLMs has provoked interest in specialist vector databases while vendors of existing databases are working to add support to bring vector search to data already stored in their data platforms. SingleStore has supported vectors since 2017 and has recently been talking up the advantages of using an existing database for vector similarity search.

SingleStore is competing with some of the biggest names in data platforms and data processing, and adoption relies on organizations understanding the availability of alternatives to traditional approaches that split data processing between operational and analytic data platforms. Industry trends are playing to the strengths of the company and its support for hybrid data processing and multi-format data storage. I recommend that organizations assessing database providers for new application development and deployment projects include SingleStore in their evaluations.

Regards,

Matt Aslett