Discover how the decoupled compute and storage architecture of the Firebolt data warehouse enables sub-second query performance on terabyte-scale data sets.
The following diagram gives a high-level overview of Firebolt’s structural architecture. Firebolt’s components, which include management, compute, and storage layers, interact with a variety of workloads to enhance performance, scalability, and resource efficiency.
Firebolt’s management layer handles key administrative functions including managing metadata, security settings, and observability, all in one place. Through this layer, administrators can oversee and control user access, permissions, and roles, ensuring robust security. It also provides insights into system performance and resource usage, providing operational visibility. The management layer also streamlines workspace management, enabling seamless organization and monitoring of data environments, further simplifying the overall administration process.
Firebolt’s management layer consists of the following:
Firebolt’s compute layer is responsible for running queries and processing data through its scalable engines. Engines use parallel processing to deliver high performance and efficiency. You can create and configure multiple engines tailored to different workflows, such as data integration or analytical queries supporting customer facing analytics. Engines can be configured for different needs like query latency, throughput, and concurrency, adapting to specific data processing tasks.
Firebolt’s compute layer features the following:
Firebolt’s storage layer efficiently manages large amounts of data by keeping storage separate from the compute process, which means that you can store as much data as needed without impacting compute resources. It uses cloud-based storage for high availability and durability, while also reducing costs. Data is stored in a compressed, column-based format to save space and improve query performance. Firebolt’s indexing features further speed up data retrieval by reducing the need to scan large datasets. This separation between storage and compute allows users to scale their storage needs independently, making resource management more flexible and cost-efficient.
Firebolt’s storage layer features the following:
Discover how the decoupled compute and storage architecture of the Firebolt data warehouse enables sub-second query performance on terabyte-scale data sets.
The following diagram gives a high-level overview of Firebolt’s structural architecture. Firebolt’s components, which include management, compute, and storage layers, interact with a variety of workloads to enhance performance, scalability, and resource efficiency.
Firebolt’s management layer handles key administrative functions including managing metadata, security settings, and observability, all in one place. Through this layer, administrators can oversee and control user access, permissions, and roles, ensuring robust security. It also provides insights into system performance and resource usage, providing operational visibility. The management layer also streamlines workspace management, enabling seamless organization and monitoring of data environments, further simplifying the overall administration process.
Firebolt’s management layer consists of the following:
Firebolt’s compute layer is responsible for running queries and processing data through its scalable engines. Engines use parallel processing to deliver high performance and efficiency. You can create and configure multiple engines tailored to different workflows, such as data integration or analytical queries supporting customer facing analytics. Engines can be configured for different needs like query latency, throughput, and concurrency, adapting to specific data processing tasks.
Firebolt’s compute layer features the following:
Firebolt’s storage layer efficiently manages large amounts of data by keeping storage separate from the compute process, which means that you can store as much data as needed without impacting compute resources. It uses cloud-based storage for high availability and durability, while also reducing costs. Data is stored in a compressed, column-based format to save space and improve query performance. Firebolt’s indexing features further speed up data retrieval by reducing the need to scan large datasets. This separation between storage and compute allows users to scale their storage needs independently, making resource management more flexible and cost-efficient.
Firebolt’s storage layer features the following: