Introducing Build to Destination (B2D)
  • 13 Jun 2022
  • 3 Minutes to read
  • Dark

Introducing Build to Destination (B2D)

  • Dark

Build to Destination (B2D) allows data modelers to choose additional destinations to Elasticubes. Instead of generating an Elasticube in Sisense, the defined data is generated in the chosen CDWH. Once a build is triggered, the process and destination tables utilize the CDWH capabilities and generate the tables on the CDWH. All queries pointing to the B2D data model behave as a Live Connection to the CDWH.

The B2D capability aligns to customer needs for agile and flexible data management. New requirements and constraints keep arising on a frequent basis, this is where a single stack tool is required to support these needs. Examples for such cases are Governance restrictions, immediate scale needs, simple migration to Live Connection, PoC, Materialization, Self Service and more. All of this is achieved using the known Data Model Interface.

B2D exploits the known build capabilities and leverages them with additional features that are designed to work well with the designated CDWH provider, such as Build with Upsert, Extended JDBC data types support and the ability to choose the preferred SQL dialect in Custom Table/Column.

B2D Benefits

Update Insert (Upsert)

Note: Currently, this is only supported in Snowflake. In addition to the current build type options available at the Table level (Full, Append, Accumulate, Changes Only), data modelers can choose to build with Update Insert (Upsert). This additional strategy aligns with the growing customer need to scale and better manage their data, while improving build times.  See Select Build Behavior in Working with Build to Destination.

Improved Build Flow

Build Duration is optimized on each consequent build. This is achieved by prioritizing the process stages and optimizing the parallelism process.

Short Circuit Optimization

In the use case when the source and destination are on the same CDWH, the data is copied locally inside the CDWH and does not pass through the Sisense import process (parquet file, upload, bulk insert). This flow is designed in this manner to remove I/O and Network redundant steps, and hence to improve build performance.

Supported Data Types in Model

B2D supports 17 JDBC data types in the Data Model. This applies to Source (Imported) Tables, Custom Columns and Custom Tables. A list can be found here.

Custom Tables and Custom Columns Process on CDWH

The Custom Columns and Custom Tables processing is done and applied in the CDWH. As with the Elasticube build process, you first generate the required tables, and then generate the required columns and tables in the CDWH. This additional optimization feature allows faster build times by reducing I/O processing time and utilizing the CDWH strengths.

Different Connectivity Configurations for Query and ETL

There are two major benefits for using different connections:

  1. Security - Making sure the Query Model and the Processing Model use different connections ensures that viewers and dashboard designers are not able to use DML commands on the B2D Data Source.
  2. Cost Saving - CDWHs that support independent physical DWH storage and processing engine, enabling the use of the same DWH Storage with different processing engines. The ability to introduce a different processing engine utilizes customer costs.

Flexible Schema Selection

Data modelers can pick the designated schema as part of the destination in the B2D data model. This allows the data modeler to work with different schemas in different data models. The schema must be a new one, and not an existing one in the destination. Each cube must have a different schema.

Optimized BI Database

Utilizing the Build capabilities ensures there are no interruptions to queries during the build process. A different data set is generated, processed and replaced once completed.

Requirements for B2D

  • Sisense version L2022.5 onwards
  • Supported destinations: Snowflake and Redshift
  • Requires a CDWH and Amazon S3 Storage with Write permissions to a bucket
  • Sisense Admin must configure the destination environment. Further details found here.
  • See a more detailed list of prerequisites here.

Was this article helpful?