GPU-Accelerated Data Warehouse

Built ground-up for big data

SQream DB was built from scratch to empower data consumers

Learn more →

Flexible at any scale

100x faster, with independent storage and compute scaling

Learn more →

Intergrate with anything

Connect any tool to SQream DB with ODBC, JDBC, and a variety of APIs

Learn more →

SQream DB - Built for scale

SQream was founded to address frustration with existing data warehousing systems, which resulted in building the first enterprise-grade GPU accelerated data warehouse. We call this product SQream DB.

Rather than building on unsuitable technology stacks like Hadoop or Postgres, SQream DB was built from scratch to empower data consumers. SQream DB was built to harness the raw brute-force power and high throughput capabilities of the GPU, with MPP-on-chip capabilities and a fully relational SQL database.

SQream DB was designed for dynamic, modern workloads that change often. We built our algorithms to handle worst-case scenarios, and we optimized for huge datasets, where typical database optimizations struggle.

Built ground-up for big data

SQream DB is a fully-featured, enterprise-ready data warehouse for analytics. It is not not an in-memory database, or an SQL translation layer for Hadoop. It is its own data warehouse, designed for larger-than-memory, constantly growing data.

Interface and Compiler

SQream DB’s interface layer is collection of services that control the data warehouse. It includes concurrency control, access control (object-level permissions), cluster management, and the statement compiler.

Compute and Execution

The compute layer is where the actual data processing tasks are run. This includes CPU and GPU resources, controlled by the SQream DB software.

Persistent Storage

The storage layer is split into the metadata layer where all routine database objects are stored, and the persistent columnar bulk data layer which is heavily optimized for raw tablescan performance.

Answers to any question, fast

SQream DB automatic tuning is a key enabler to analyzing data without intermediate steps. The raw, brute power of the GPU allows SQream DB to analyze data immediately after load. This happens during the ingest process, requiring no user intervention.

SQream DB can also read data directly from external sources, with the external table syntax, which avoids loading data before it is needed.

Spotlight on the SQream DB table

SQream DB’s optimized columnar storage system is partitioned both horizontally and vertically for best performance for heavy analytic operations like joins, aggregations, summarizations, and sorting.

Flexible and scalable

Every SQream DB instance can be thought of as an MPP database by itself, with shared-data architecture. Each instance has full access to all data in the persistent storage layer. Permissions are managed by the service layer above the SQream DB instance, based on the user’s role. These instances can be launched or shut down at will, as the requirements inevitably changes.

See more SQream DB Features

Columnar engine

This feature allows selective access to the required subset of columns, reducing disk scan and memory I/O when compared with standard row storage. This concept is well-suited for parallelized compute, like the GPU.

Hyperpartitioning - chunks and extents

SQream DB automatically splits up the storage horizontally into manageable chunks enabling efficient usage of the hardware resources and relatively small GRAM (GPU RAM) availability in GPUs. The clever use of spooling and caching help make the most of the limited GRAM.

Use the SQL you know

SQream DB supports an ANSI SQL compliant syntax, and easily integrates into existing ecosystems by supporting industry standard ODBC and JDBC connectors, as well as Python and C# .Net, C++, Java, and others.

SQream DB’s native SQL interface eases transition from other databases. There’s no need to maintain odd APIs and custom Scala code. Full SQL support lets any existing ETL and applications to connect and offload heavy database operations to SQream DB, minimizing the time needed to get up and running with the new platform.

See more SQream DB Features

Intergrate with anything

SQream DB’s automatic optimizations let businesses focus on data, rather than management

SQream DB runs on standard Linux hardware, so it's easily monitored and provisioned

Take comfort with SQream DB's object-level permission system. Control roles and objects all the way down to per-table authentication.

“SQream helps us to keep pace with rapidly increasing data usage and translate that data into real benefits for our customers, whether in helping to manage the quality of our networks or enabling us to keep ahead of our competition.”
Suppachai Panichayunon
Head of Solution Design and Architect, AIS
“With our aim to be future ready, SQream helps us reduce the query time and create value for our customers.”
Ajit Singh
Sr. VP Engineering, ACL Mobile
“We saw a tremendously cost effective opportunity to obtain comprehensive analytic abilities we didn’t have before SQream, required to continuously improve our network service for our customers.”
RF Group Leader
Cellcom