Performance & Scalability
Last updated
Last updated
Platform supports Multi-tenancy. User License is flexible and its based on the Instance and not on the tenant. For 100 concurrent users the following infrastructure is required 24 Cores & 64 GB Memory. Infrastructure can be scaled up to support more concurrent users.
The BDB Platform is deployed on Kubernetes and based on microservice architecture. Kubernetes allows the auto-scaling if it is enabled while deploying the platform.
A dashboard with 6 ClickHouse data service calls and 7 chart components and many label components, took around 4 seconds to render in the browser. The queries are processing over ClickHouse DB tables approximately with 1.5 Million records (to the max). The service response on average is between 1 to 2 seconds and the max number of rows returned is 1000 rows with 7 columns.
More details on the Performance Report can be viewed at the - https://bdb.ai/docs-performance
The BDB Platform doesn’t use any in-memory cache for Dashboards. All the services are passed through queries that get executed at the database level. Hence, the data handling capacity is based on the underline Database infrastructure and configuration.
The BDB Dashboards can consume real-time data streaming via webhook. Automatic data updates can be done via a pre-configured auto refresh facility based on event or time.
The BDB Dashboard Designer can be configured to automatically refresh a dashboard. The dashboard data services can be configured to refresh promptly at a lower frequency of 15 seconds. Webhook connectivity will provide a real-time data refresh, it can also automatically trigger other services based on the configuration.
The BDB Platform can disable data export from a dashboard. Please explain more about what you meant by the "run-away" reports! If you have shared a Dashboard with an external user, then that control can be taken away anytime.
The BDB Platform provides an option to configure the max rows returned for all the data connectors to ensure a good user experience. The data services support pagination. The BDB Dashboard Designer contains built-in data grid components with pagination.
There is no specific limit to the number of tenants or users in the multi-tenant cloud deployment. As a best practice, the BDB team recommends setting 2000 tenants per instance.
BDB Conducts a pen test by third-party annually for major releases. PEN test reports can be shared with the customer based on specific requests. Performance Report can be viewed at - https://bdb.ai/docs-performance.
The Performance test reports can be viewed at: https://bdb.ai/docs-performance
The actual performance experienced by our customer is depended on various factors. Such as,
the size and complexity of the dataset,
the hardware and infrastructure being used,
the specific visualizations and calculations being performed
The BDB Platform is capable of providing quick and effective insights from non-AI models with appropriate hardware and optimization.
The performance of AI models within the BDB Platform can vary greatly depending on the specific use case and the complexity of the AI models being used. In general, the performance of AI models within the BDB Platform will depend on factors such as,
the size and complexity of the dataset
the performance of the hardware and infrastructure being used
the specific AI algorithms and models being employed
In general, customers should expect some level of performance trade-off when using AI models within the BDB Platform, as the calculations required for AI models can be computationally intensive. However, with the right hardware and optimization, customers can still expect to get valuable insights from AI models within the BDB Platform.
The BDB Platform provides a number of performance tuning capabilities to help users optimize the performance of their data visualizations and analyses. Some of these include:
Data source optimization: The BDB Platform provides some data source optimization techniques to help users improve performance, such as data aggregation, indexing, and materialized views.
Data visualization optimization: The BDB Platform provides several techniques for optimizing the performance of data visualizations, including best practices for chart types, data densification, and level of detail expressions.
Hardware optimization: The BDB Platform allows users to take advantage of high-performance hardware, such as graphics processing units (GPUs) and solid-state drives (SSDs), to improve performance.
Server and network optimization: The BDB Platform provides the server and network optimization techniques, such as load balancing and data compression, to help users improve the performance of their deployments.
Data blending and aggregation optimization: The BDB Platform provides a number of techniques for optimizing the performance of data blending and aggregation, including data extract optimization and query optimization. These are just a few of the performance-tuning capabilities available to users of the BDB platform. By taking advantage of these capabilities, users can help ensure that their BDB Platform deployments provide the fast and responsive performance they need to effectively explore and visualize their data.
