Familiarizing the BI tool architecture takes you one step further in getting to know better about your BI tools and Capabilities/Offering. Capacity planning for the new implementation still requires basic Hardware and Software experience and Understanding along with the suggestions from the Product Vendor.
In my quest of 2nd week working for Tableau/Reporting support, I started reading this Whitepaper from Tableau - https://www.tableau.com/sites/default/files/whitepapers/whitepaper_tableau-for-the-enterprise_0.pdf.
If you are part of any BI product installation or familiar with architecture, learning Tableau relating to other BI products can be relatively interesting.
Tableau is N-tier architecture with exciting in-memory capability.
Following are the layers of Tableau Architecture.
- Customer Data - Heterogeneous source systems
- Data Connectors - Fast Data Engine/Native Connectors
- Main Components - Data/VizQL/Application
- Gateway - Gateway/Load Balancer
- Clients - Desktop/Mobile/Web
Unique offering of Data Layer - I feel Business and IT organizations of the enterprises worldwide pushed the Analytics Vendors to have this capability of combining the heterogeneous source systems (I know what you are thinking, streamlining business process to make use of traditional DW and BI tools can be hard especially when business users have many more excel files to be analyzed.) Explosion of evolving potential Data sources in the day to day business requires IT and Business to be agile and nimble, Tableau and the likes were meant for such business users/process wherein you are requested to integrate many of these data sources with your existing Data warehouse/Decision support systems. Unique offering with Tableau data layer is it can leverage the power of Database engine for the analysis and also on the other hand support the power users with the in-memory Data engine which improves the speed of the data analysis to be in-par with the expectation.
Data Connectors - I like the mode of connection capability with respect to the data source connectivity, Provides ability to report on live data and also in-memory mode to leverage the in-memory data engine.
Especially the below point seems to be interesting, Would like to experiment with some large data extracts. "Because the Data Engine can access disk storage as well as RAM and cache memory, it is not limited by the amount of memory on a system. There is no requirement that an entire data set be loaded into memory to achieve its performance goals."
Tableau Server Components - Tableau server components includes 4 processes.
Application Server (wgserver.exe) - handles browsing and permission.
VizQL Server (vizqlserver.exe) - Handles the data query to Data source which is rendered in desired format. Includes a cache which can be shared across multiple users.
Data Server - centrally manage and store Tableau data sources. It also maintains metadata from Tableau Desktop, such as calculations, definitions, and groups.
Backgrounder - The backgrounder refreshes scheduled extracts and manages other background tasks.
Gateway/ Load Balancer - Similar to the other BI tools, Gateway handles the end user request and assign to the appropriate process and distributes the load. If you have worked on Cognos installation/environments, this also does the same.
Tableau Client Interfaces - Tableau Server provides interactive dashboards to users via zero-footprint HTML and JavaScript (AJAX) in a web browser, or natively via a mobile app.
Tableau Desktop - rapid-fire authoring environment used to create and publish views, reports and dashboards to Tableau Server. Also can access the published views/data sources on the server.
Happy Learning!
