December 2019 Summaries

The SME Solutions Group, in collaboration with SingleStore, helped a utility company overcome its complex and outdated data infrastructure by replacing 10 different components with a single SingleStore cluster. The new solution provided outstanding performance, scalability, and the ability to use standard business intelligence tools via SQL, while reducing costs. The utility company's data ingestion requirements were met as it streamed in 100,000 rows per second, and the system could handle aggregations on three levels, bucketing reads into high alarms and low alarms, querying data throughout the process, and maintaining performance with tight service level agreements. The SingleStore solution also enabled machine learning and AI capabilities, making it easier to implement and unify data for comparison, queries, and more complex operations. With SingleStore at its core, the utility company can now run analytics and reporting across their entire dataset using various tools and ad-hoc processes, scaling easily as their data grows to billions of rows.

SingleStore, a relational database, utilizes skiplist indexes instead of traditional B-tree or similar structures for in-memory data. This choice is driven by the need for memory-optimized and simple indexing that can be easily implemented in a lock-free fashion, making it fast and flexible. The use of skiplists eliminates the overhead associated with indirection, reducing the number of instructions required to insert, delete, search, or iterate over data. In contrast, traditional databases like MySQL rely on more complex locking schemes and have higher memory overhead due to page splitting and fragmentation. SingleStore's implementation of skiplist indexes results in significant performance improvements, particularly when compared to B-tree implementations like those found in SQL Server and InnoDB, with advantages extending to concurrent write workloads and reverse iteration support.

SingleStore is a high-performance, scalable SQL relational database system that supports operational analytics and has capabilities for predictive ML and AI. It's designed to handle large volumes of time series data with ease, providing fast ingest, powerful query processing, and scalability through scale-out. The new Self-Managed 7.0 release introduces special-purpose features making it even easier to manage time series data, including a brief syntax for querying, built-in functions like FIRST(), LAST(), and TIME_BUCKET(), and support for series timestamp columns. This allows non-expert users to perform common queries on time series data more easily and makes expert users more productive. The database supports multi-tenancy, encryption, data access roles, management, row-level security, and has utilities available for time series data migration from existing databases like Informix. SingleStore is a cloud-native platform that can be run in various environments, including on-premises, in the cloud as a managed database platform, or as a service with Singlestore Helios.

SingleStoreDB Self-Managed 7.0 uniquely suits real-time analytics with its scaled-out architecture, fast query execution, and non-blocking concurrency control. The new release introduces special-purpose features to make time-series data management easier, including FIRST(), LAST(), TIME_BUCKET(), and the ability to designate a table column as the SERIES TIMESTAMP. These features allow for concise queries using fewer lines of code and complex concepts, making it more accessible to developers who manage time series data. The new functions enable simple SQL operations on time series data by providing an implicit argument for time-based functions, reducing the need for explicit referencing of the time attribute in every query expression related to time. With these enhancements, SingleStoreDB Self-Managed 7.0 provides better overall data management support for customers who manage time series data, offering performance, scalability, reliability, SQL support, extensibility, and rich data type support.

SingleStore has announced the general availability of its database management system, SingleStoreDB Self-Managed 7.0, which is now available on-demand with its elastic cloud database, Singlestore Helios. This new release delivers breakthrough capabilities for operational analytics, machine learning and AI, including enhanced resilience features, table type convergence, time series enhancements, and more. With the public availability of this new version, SingleStore further solidifies itself as a powerful fit for companies' most critical operational workloads, offering effortless deployment, elastic scalability, enhanced ease of use, reduced total cost of ownership, and greatly increased flexibility. The combined advantages of Singlestore Helios and SingleStoreDB Self-Managed 7.0 become available on AWS and other cloud platforms, providing users with the full capabilities of one-click deployment and easy cloud scalability.