Integration Services Programming Architecture

New: 14 April 2006

Integration Services has a completely new architecture that separates data movement and transformation from package control flow and management. There are two distinct engines that define this architecture and that can be automated and extended when programming Integration Services. The run-time engine implements the control flow and package management infrastructure that lets developers control the flow of execution and set options for logging, event handlers, and variables. The data flow engine is a specialized, high performance engine that is exclusively dedicated to extracting, transforming, and loading data. When programming Integration Services, you will be programming against these two engines.

The following image depicts the architecture of Integration Services.

Integration Services Run-time Engine

The Integration Services run-time engine controls the management and execution of packages, by implementing the infrastructure that enables execution order, logging, variables, and event handling. Programming the Integration Services run-time engine lets developers automate the creation, configuration, and execution of packages and create custom tasks and other extensions.

For more information, see Extending the Package with the Script Task, Developing a Custom Task, and Building Packages Programmatically.

Integration Services Data Flow Engine

The data flow engine manages the data flow task, which is a specialized, high performance task dedicated to moving and transforming data from disparate sources. Unlike other tasks, the data flow task contains additional objects called data flow components, which can be sources, transformations, or destinations. These components are the core moving parts of the task. They define the movement and transformation of data. Programming the data flow engine lets developers automate the creation and configuration of the components in a data flow task, and create custom components.

For more information, see Extending the Data Flow with the Script Component, Developing a Custom Data Flow Component, and Building Packages Programmatically.

Supported Languages

SQL Server 2005 Integration Services fully supports the Microsoft .NET Framework. This lets developers program Integration Services in their choice of .NET-compliant languages. Although both the run-time engine and the data flow engine are written in native code, they are both available through a fully managed object model.

You can program Integration Services packages, custom tasks, and components in Microsoft Visual Studio .NET or in another code or text editor. Visual Studio .NET offers the developer many tools and features to simplify and accelerate the iterative cycles of coding, debugging, and testing. Visual Studio .NET also makes deployment easier. However, you do not need Visual Studio .NET to compile and build Integration Services code projects. The Microsoft .NET Framework SDK includes the Visual Basic .NET and C# compilers and related tools. For more information about how to compile code projects using the command-line compilers, see Installing and Compiling Integration Services Programming Samples.

The Integration Services Script task and Script component use Microsoft Visual Studio for Applications (VSA) as an embedded scripting environment. VSA supports only Visual Basic .NET as a programming language.

Commonly Used Assemblies

The following table lists the assemblies that are frequently used when programming Integration Services using the .NET Framework.

Change History