New BizTalk Server Community Forum

New BizTalk Server Community Forum has launched a new BizTalk Community Forum! 

Biztalk Forum

This is not intended to replace the Microsoft Official Newsgroups.  That should always be your first place to turn to for technical questions.  But, if you are looking for a smaller, more relaxed area check it out.

No registration is required to vote in polls, view threads, and post to the forum.

Registered users can get forum level notifications, topic reply notifications, and create a user profile.

Forum categories include:
Site and Blog Comments
BizTalk Server 2004
BizTalk Server 2006
Windows Workflow Foundation
Job Postings (both full time and contract)

If you have questions / comments about any of my blog posts or samples, the forum is a great place to ask them!

Do not forget to check out our weekly polls.


Most recent polls:
What type of messages do you typically process with BizTalk?
Do you use other tools for transformations other than the Biztalk Mapper?


Do not have time to visit the forum?  Get the 50 most recent posts via RSS.

BizTalk Server 2004: Further observations on Orchestration message handling

Richard Seroter of Microsoft has published a useful blog posting at in which he describes one technique for passing messages of any type through a BizTalk orchestration.   Here is some further information to supplement what he has written.

A Messaging-based State Machine Design Pattern, using BizTalk Server

Here’s something I’ve been meaning to write up for a long time. Originally intended as a blog post, it turned into an article for .NET Developer’s Journal, and now, at last, here it is on my blog. Thanks to MartyWaz for first enlightening me to the technique on a long-ago POC project.

The intended audience is an intermediate/advanced BizTalk person. If you’re a beginner, that’s fine, but if you don’t understand it, read it again. This is a VERY cool way to create BizTalk solutions.



Decoupling BizTalk Orchestration Processes by Using Direct Binding

As all architects and developers know, the tenets of service-oriented architecture call for breaking large monolithic processes into more granular, purpose-specific blocks of functionality that solve specific needs, and exposing those as services. This is not really new thinking. Languages have long supported the notion of breaking logic into discrete units. If applied properly, this approach will yield a series of services that can potentially be aggregated in different ways to provide different solutions. In short, this building-the-building-blocks approach is a cornerstone of reuse. In contemporary development trends, these chunks of functionality are increasingly exposed as Web services.

From a process-engineering standpoint, it makes sense to apply the same approach to the business processes themselves, and for the same reasons. If you decompose a large monolithic business process into more granular subprocesses, then you increase the chances of being able to reuse those subprocesses.

As an example of this, consider the new hire process, an often complex process that could include tasks such as creating network credentials, creating mailboxes, registering with human resources and payroll, provisioning telephones, ordering business cards, etc. This process could be automated as one large monolithic service. Using BizTalk Server 2004, this “single process” view could be implemented as one large orchestration that sequentially, or in a parallel (or some combination thereof) invoked a series of steps to implement this:

While perfectly viable, such approaches would lead to a monolithic block, with no chance of being able to easily reaggregate the composite services. In our example, business cards will need to be reordered, phones may need to be reprovisioned. We may do some “copy and paste” reuse, but that’s the best we could hope for – and that’s just not good enough in the current agile business environment.

A more granular approach could have a series of orchestrations calling each other; for example, first orchestration sets up network credentials, calls the second one that orders business cards, which calls the third that provisions telephones, and so on. The drawback of this approach is that everything is hard coded, such as the next step in the process, and it is difficult to inject other steps into the process, as well as impossible to reuse a step as part of a different process. Hence, any new subprocess would require a code change and redeployment

A third approach would be to have a “master” or “controller” orchestration that invokes the subprocess orchestrations, controlling their execution and order of execution. While it’s easier to add or remove processes, this approach would still require code changes and redeployment.

The purpose of this article is to show you an alternate approach that promotes granularization, while also offering significant benefits in terms of scalability, agility, and extensibility.

Loosely Coupled Solution
BizTalk includes a powerful publish/subscribe mechanism. Most developers today have backgrounds in code-oriented, procedural, or object-oriented development paradigms, and as such when people start developing BizTalk solutions they often overlook BizTalk’s message-oriented capabilities. .

The publish/subscribe mechanism works by creating and filling subscriptions. When a new message arrives in the MessageBox, a message agent looks for subscribers, and sends the message to any endpoints that have subscriptions. Subscriptions could have been set up in several ways, including by binding an orchestration to a receive port, having a correlated receive waiting for a message, creating a send port with a filter condition that matches something about the message (type, receive point, value of a routable property, etc.), and so on.

This is an efficient and scalable approach. Wouldn’t it be nice if we could harness the power of the publish/subscribe mechanism to bind together our subprocesses? If we did that, then we could create a series of discrete subprocesses, define the types of messages that trigger their invocation, and not be worried about the sequence. A process would be activated by the appearance of a message, do its work, then perhaps drop a message back into the MessageBox that could in turn activate another subprocess (or multiple subprocesses). The good news is that we can do just that.

This could work in several ways, but for the purposes of this example let’s assume that we want the processes to execute sequentially. Since that is the case, a simple way to implement this would be to have a single “new hire” message that each process would work with (although these could of course also be completely different messages). This new hire message could have a Status field, which would be a promoted property. By promoting it, the Status field would become a routable field, which makes it visible to the publish/subscribe mechanism. You could define the execution order of the processes by filtering the orchestration activating Receives by a specific status value. In our example, the NetworkCredentials activating Receive picks up a new employee request message from the file system, assigns “CREDENTIALS_DONE” to the Status field, and sends it to the MessageBox through a direct bound port. The OrderBusinessCards orchestration’s activating Receive shape’s filter condition is watching for a status of “CREDENTIALS_DONE” as shown below:

For the receive ports in the orchestrations, we specify that they are direct bound: .

When the OrderBusinessCards orchestration is done, it will update the Status field with “BIZCARDS_DONE” and persist the message to the Message Box. This is done by having a send port in the orchestration that is direct bound to the Message Box (just as we specified for our receive, but this time in the send direction).

The final orchestration to run will be the ProvisionTelephone orchestration, which will be filtering for messages that have a Status of “BIZCARDS_DONE,” and will update the Status field with “TELEPHONE_DONE” and persist the message to the file system.

Extending the Solution
What has been shown thus far is a business process that consists of three orchestrations that are loosely coupled through the Message Box. By using this approach, our process can robustly handle large peaks in message volumes without becoming overwhelmed. Want to scale this to a half-million messages per hour? Sure, we’ll need some more hardware, but it wouldn’t be a problem. .

What if we wanted to extend the process? Say for example that we now want to implement an audit trail, where a request would be persisted whenever new network credentials are created. No problem at all. Just create a new send port that uses the file transport to deposit the message in the file system, set a filter condition on the send port, and start it up. By using this approach, without making any code changes to the existing (and deployed) process, we have now inserted another step into the new hire process.

Note that when you specify string values in a send port filter, the string is not in quotes. If you specify it in a receive shape’s filter, the value must be in quotes.

In our example we have opted to run the three processes in sequential order, but we did not need to. The NetworkCredentials orchestration could have run updating the Status with “CREDENTIALS_DONE,” the OrderBusinessCards and ProvisionTelephones could both have been filtering for a message with a Status of “CREDENTIALS_DONE,” and they would have run in parallel. With the modular and decoupled design presented herein, you see how easy it is to change the sequence of subprocess execution.

Sample Code
The sample code for this article is available here.

To run it:

  • Unzip it using folder names.
  • The project was originally located under c:\Visual Studio Projects. If your pathing is different, you will need to edit the binding.xml file to set appropriate paths for the FILEDROP folder so we can pick up messages from the file system and persist them.
  • Build and deploy the solution.
  • Import the binding.xml file.
  • Start the DecoupledProcess.* orchestrations.
  • Copy the NewEmpReq_Instance.xml document to the filedrop\in folder.
  • See the output message appear in the Out_ProcessCompleted folder.
BizTalk is a powerful tool that allows developers to rapidly create comprehensive integration and business solutions. As is the case with any development tool or framework, for any given requirement, there are usually several ways to create solutions. This article has demonstrated how you can use a message-oriented paradigm, leveraging BizTalk’s publish/subscribe mechanism, to create scalable and agile solutions.

Of CTP bits and South Africa

Several people have asked me whether there will be a build of BizTalk Server that works on the RTM bits of SQL Server and Visual Studio .NET for the Launch.  That’s the plan.  It won’t be beta 2 because we are still working on our quality metrics versus beta 2 however it will be pretty solid and if you are really lucky we will release it on November 7th.  Indeed, we are hopeful that you might get both the BizTalk CTP and another CTP all on November 7th. More on that later.

I’m currently in Sun City, South Africa.  Every year I have a goal to visit a particular place and this year I really wanted to visit South Africa.  With  my current role I’m very lucky to be able to combine work and pleasure.  I’ve just presented a couple of sessions on BizTalk and Workflow at TechEd South Africa and in a day or so I’ll put the computer to aside and go exploring in Kruger and up in Zambia (Victoria Falls) which I’m super looking forward to. 

I really need to recharge because the rest of the month is going to be crazy fun with the BTS/SQL/VS launches.  I’ll be keynoting the BTS/SQL/VS Launch event in Detroit on November 8th and a similar event on November 29th in Montreal so if you live in either of those areas come along and say hi.

Btw it seems a few people are confused about WinFX.  WinFX is the managed programming API for Windows (version independent) just like Win32 is unmanaged programming API for Windows. 

Then there is the WinFX Runtime Components 3.0 which is the runtime redistributable package is versioned and available for platforms that require it to be installed.  WinFX Runtime Components 3.0 includes the stuff that used to be called the .NET Framework, Windows Presentation Foundation, Windows Communication Foundation, and of course Windows Workflow Foundation. 

Processing a Large Flat File Message with BizTalk and the SqlBulkInsert Adapter

Below discusses the second demo for a presentation that I did at the Business Process Integration & Workflow Conference in Redmond during the week of Oct 4.

The title of the presentation was entitled:
Handling Large Messages in BizTalk

The first demo that I did for the presentation can be found here:
Processing a large message and a faster, less CPU intensive splitter pattern.

This entry compares two different methods for processing large flat files in BizTalk.
The size of the flat files being tested ranged in size from: 8MB to 201MB.

At the end of this entry is the download for the SqlBulkInsert Adapter code and the BizTalk Test Project to
perform the tests.

A section of the flat file message being processed is as below. About half of the columns are cut off:

The above is the flat file representation of the XML Production orders message that were used in the first Demo.
Each line of the flat file message contains a production order. The production order data (columns) are separated by commas and each production order line (record) is separated by a Carriage Return/Line Feed.

To process the flat file message, two different implementations were used. In both implementations, the flat file is pre-processed so that it is in a format for further processing -> For example mapping or splitting.

Implementation One: Processing with a Flat File Schema and Custom Receive Pipeline

This implementation uses the standard method of using a XSD Schema with flat file extensions to represent the above flat file message. A custom receive pipeline utilizing a flat file disassembler transforms the flat file into an internal XML message. This XML message will be published into the MessageBox database. If you have not processed a flat file in BizTalk before, try this example in the SDK.

This is just a short recap of how this implementation works and is really for the benefit for those not familiar with BizTalk or who have never processed a flat file in BizTalk.

a) Flat File is picked up by a Receive Location.
b) Receive Location is configured to use a Custom Receive Pipeline
c) Custom Receive Pipeline will convert (disassemble) the flat file, into an internal XML message as below:

d) This XML Message is then published into the MessageBox database
e) Once the XML Message is published into the MessageBox database, any number of Orchestrations or send Ports can subscribe to this message and process this message. This message could then be mapped into another message or split into separate messages as in the first demo.

Implementation Two: Processing with the SqlBulkInsert Adapter

From a performance perspective there are three areas where the above implementation can be improved. This is especially true, if the incoming flat files messages are large greater than 20 MB.
Areas of improvement are listed below:

1) The pipeline processing time.
2) CPU utilization is at 100% for pipeline processing.
3) Flat files when converted into XML files can double or triple in size. For example when a 201MB flat file production order message is converted into an internal XML message, the size of this internal XML message expands to 767MB. This large 767MB message is then published into the MessageBox database. The size of the XML message is greater because of added tags, elements and attributes.

Therefore one approach would be to skip the pipeline processing for the large flat file message. Pipeline processing is a necessary and great feature in BizTalk, but will be skipped in this implementation to increase performance.

A Custom Adapter (SqlBulkInsert Adapter) will be used to pre-process the flat file.
This adapter can only be used with a One-Way Receive Port. The property pages of the Receive Location are as below:

How this implementation works:

a) This adapter behaves similar to the out of the box File Adapter used in a Receive Location. It will poll a directory on a Hard Drive for files to be picked up and processed.
b) The FileReceiveFolder Property is set to a directory on a hard drive.
For example -> C:\BTSBulkLoad\Files\ReceiveForSqlBulkLoad
Flat Files dropped into this folder will be picked up to be processed.

c) The FileMask Property is set so the adapter will only pick up only files with the mask.
For example -> *.txt

d) The FileErrorFolder Property will be populated with error files.
For example -> C:\BTSBulkLoad\Files\FileErrorFolder.
If some of the rows(data) in the flat file message are malformed, these will not be processed , but will be placed in a file for viewing. This is analogous to a recoverable interchange. Rows in the flat file that are correct in structure will be processed.

For example, if two of the rows in the flat file cannot be processed (Missing Columns, missing delimiters etc) a file is produced with the row(s) that could not be processed.

11506,10020,TST GF 01 GRD 01,7045936818561,20.25,20.25,58.25,58.250359010995972,3 Inch Core,,,
11665,10020,TST GF 01 GRD 01,36,23584,23584,1432.7045936818561,1432.7045936818561,20.25,20.25,5

Additionally another file is produced with a more detailed explanation of why the row(s) could not be processed as below:

Row 16 File Offset 3822 ErrorFile Offset 0 – HRESULT 0x80004005
Row 25 File Offset 6422 ErrorFile Offset 493 – HRESULT 0x80004005

The Max Number of rows with errors can be configured (explained below). If this number is exceeded, the operation will fail as a whole.

Once a File has been picked up, the SqlBulkInsert adapter will use the below properties to delegate the processing of the flat file to a Sql server stored procedure. The large File will not be submitted to the BizTalk engine to be processed. Therefore the BizTalk pipeline processing will be skipped for the large file.

e) The SqlConnectionString Property is set with the connection string to the sql database that will process the flat file.
For example -> packet size=4096;integrated security=SSPI;data source=”(local)”;persist security info=False;initial catalog=BTSSqlBulkLoad 

f) The SqlStoredProcedureName Property is set to the name of the stored procedure that will process the flat file.
For example -> LoadInsertProductionOrders

The declaration of the stored procedure is as below:

CREATE Procedure [dbo].[LoadInsertProductionOrders]

@pathAndFileName  varchar(500),
@pathAndErrorFileName varchar(500)

g) The SqlStoredProcFileNameParameter property will set with the name of  the parameter that accepts the file to be processed.
For example -> @pathAndFileName

At run time when the adapter picks up the file, it will generate the following to populate the @pathAndFileName stored procedure parameter:
For example -> C:\BTSBulkLoad\Files\ReceiveForSqlBulkLoad\FlatFileToProcess.txt

Note: Only a pointer to the flat file to be processed is passed to the stored procedure.
Note: UNC paths should also work, but have not actually tried.

h) The SqlStoredProcFileErrorParameter property is set to the name of the parameter in the stored procedure
to indicate the location to where the error files should be placed. See part d)
For example -> @pathAndErrorFileName

At run time when the adapter picks up the file, it will set the @pathAndErrorFileName stored procedure parameter to a Value.
For example -> C:\BTSBulkLoad\Files\FileErrorFolder

At run time when the File Receive portion of the adapter has picked up a file to process, the following code is then called in the adapter to invoke the stored procedure:

internal static string callBulkInsertStoredProcedure
   (string connectionString,
   int commandTimeOut,
   string storedProcedureName,
   string fileName,
   string fileErrorName,
   string fileNameParameterName,
   string fileErrorNameParameterName)
 System.Text.StringBuilder sbForReturnedXML = new System.Text.StringBuilder();
 SqlConnection sqlConnect = new SqlConnection();
 sqlConnect.ConnectionString = connectionString;
 SqlCommand sqlCommand = new SqlCommand();
 sqlCommand.CommandType = System.Data.CommandType.StoredProcedure;
 sqlCommand.CommandText = storedProcedureName;
 // Dynamically set the name of the parameters that should be called in the custom stored procedure.
 // This is because each stored procedure may have parameters with different names.
 sqlCommand.Parameters.Add(new System.Data.SqlClient.SqlParameter(fileNameParameterName, System.Data.SqlDbType.VarChar,500));
 sqlCommand.Parameters.Add(new System.Data.SqlClient.SqlParameter(fileErrorNameParameterName, System.Data.SqlDbType.VarChar, 500));
 sqlCommand.Parameters[fileNameParameterName].Value = fileName;
 sqlCommand.Parameters[fileErrorNameParameterName].Value = fileErrorName;
 sqlCommand.Connection = sqlConnect;
 sqlCommand.CommandTimeout = commandTimeOut;
 System.Xml.XmlReader xmlReader = null;

  xmlReader = sqlCommand.ExecuteXmlReader();
  string xmlProductionOrderIdNode;
  while (!xmlReader.EOF)
   xmlProductionOrderIdNode = xmlReader.ReadOuterXml();
   sbForReturnedXML = sbForReturnedXML.Append(xmlProductionOrderIdNode);

  catch (System.Exception ex)
   System.Diagnostics.Debug.WriteLine(“Error Occured in Adapter ” + ex.Message);
   throw ex;

  return sbForReturnedXML.ToString();

The above is just basic ADO.NET code to call the stored procedure that will return back a small
XML message generated by the stored procedure. This XML message can contain any information. This small XML message will
be published into the MessageBox database. (More about this below).

The stored procedure configured in the receive location is executed. This is where the real processing of the flat file takes place.
Note: This stored procedure could be altered in any way to process the flat file. The below is just an example.

CREATE Procedure [dbo].[LoadInsertProductionOrders]

@pathAndFileName  varchar(500),
@pathAndErrorFileName varchar(500)

— This stored procedure will take the passed path of the flat file (produced by the the adapter) to process.
— For example c:\directory\FlatFile.txt.
— The code in this procedure will then Bulk Insert this flat file into a Sql Server table.
— This stored procedure was originally created in Sql 2000, but was moved to Sql 2005
— for the following reasons:

— a) Sql 2005 has introduced a new argument for Bulk Insert Clause [ [ , ] ERRORFILE = ‘file_name’ ] 
— This will place malformed rows into an error file.

— b) Sql 2005 has introduced Try Catch Blocks in T-Sql. In Sql 2000, if there were any malformed
— rows in the flat file, an error would be raised to the .Net Helper Component and the operation would cease.
— With the Try Catch Block in Sql 2005, no error is raised to the .Net Helper Conponent and processing continues.


— Flat File rows will be inserted into this temp table.
— Note: Could of used a Format File  file, but this temp table matches the structure of the flat file.
— Eventually, the rows in this temp table will be inserted into a permanent table.


CREATE TABLE #tempInsertNewProductionOrders (
 [trk_unit_id] [int] NOT NULL ,
 [pro_product_id] [int] NOT NULL ,
 [actual_grade] [varchar] (80) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
 [actual_basis_weight] [float] NULL ,
 [actual_length_at_turnup] [float] NULL ,
 [actual_length] [float] NULL ,
 [actual_weight_at_turnup] [float] NULL ,
 [actual_weight] [float] NULL ,
 [required_width] [float] NULL ,
 [actual_width] [float] NULL ,
 [required_diameter] [float] NULL ,
 [actual_diameter] [float] NULL ,
 [actual_core] [varchar] (80) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
 [actual_property_1] [varchar] (255) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
 [actual_property_2] [varchar] (255) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
 [actual_property_3] [varchar] (255) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
 [update_timechain] [smallint] NULL ,
 [update_time] [varchar] (30) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
 [update_user_id] [int] NULL ,
 [position_index] [int] NOT NULL ,
 [comment] [varchar] (255) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
 [required_length] [float] NULL ,
 [required_weight] [float] NULL ,
 [actual_mfg_grade] [varchar] (80) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
 [actual_moisture] [float] NULL ,
 [actual_caliper] [float] NULL ,
 [actual_colour] [varchar] (80) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
 [actual_finish] [varchar] (80) COLLATE SQL_Latin1_General_CP1_CI_AS NULL ,
 [set_number] [int] NULL ,
 [position_percent] [int] NOT NULL ,
 [tare_weight] [float] NULL ,
 [user_scale_weight] [float] NULL ,
 [wire_side] [char] (3) COLLATE SQL_Latin1_General_CP1_CI_AS NOT NULL ,
 [trkc_length_adjust_type_id] [int] NULL ,
 [actual_compression] [float] NULL ,
 [actual_hardness] [float] NULL ,
 [sch_prod_order_id] [int] NULL ,
 [trk_set_item_id] [int] NULL ,
 [trk_unit_id_package] [int] NULL )

declare @return int
declare @rowcount int, @error int

declare @sqlStatement varchar(8000)

— Create the dynamic sql that contains the Bulk Insert Statement to Bulk Load the flat file into the Temp Table
— An Argument could of been used to control the max number of error rows. As below:
— MAXERRORS [ = max_errors ]
— MAXERRORS, Specifies the maximum number of errors that can occur before the bulk copy operation is canceled.
— Each row that cannot be imported by the bulk copy operation is ignored and counted as one error. If max_errors is not specified, the default is 10.

set @sqlStatement = ‘BULK INSERT #tempInsertNewProductionOrders’
set @sqlStatement = @sqlStatement +    ‘ FROM ‘ +  ””+ ltrim(rtrim(@pathAndFileName))  +””
set @sqlStatement = @sqlStatement +  ‘ WITH ‘
set @sqlStatement = @sqlStatement +     ‘  ( ‘
set @sqlStatement = @sqlStatement +  ‘ FIELDTERMINATOR = ‘  +   ”” + ‘,’ + ”” + ‘,’
set @sqlStatement = @sqlStatement  + ‘ ROWTERMINATOR = ‘ +  ””  +  ‘\n’   + ”” + ‘,’
set @sqlStatement = @sqlStatement  + ‘ ERRORFILE = ‘ +  ””  +  ltrim(rtrim(@pathAndErrorFileName ))   + ””
set @sqlStatement = @sqlStatement  +   ‘)’

— Temp Table for Sending back results

create table #results (rowsprocessed int,
           ErrorCode int,
           BatchIdentifier varchar(50))

— Create a unique GUID, that will be stored in each row.
— This is to differentiate the rows from other batches.
— This GUID, will be returned back to the Adapter in a small XML message:

declare @newid uniqueidentifier
set @newid = newid()
declare @UniqueIdentifier varchar(50)
set @UniqueIdentifier = replace(convert(varchar(50), @newid),’-‘,”)


Begin Try
 — Excecute the Bulk Insert Statement
 exec (@sqlStatement)

 — Insert the rows from the temp table into the Permanet Table.
 — For each row, also set the Batch Guid.
 Insert into InsertNewProductionOrders
 Select  #tempInsertNewProductionOrders.*,@UniqueIdentifier from #tempInsertNewProductionOrders

 Select @rowcount = @@rowcount,@error = @@error 
 Insert into #results values(@rowcount,@error,@UniqueIdentifier)
End Try
Begin Catch

    — Catch any errors and re-raise 

    DECLARE @ErrorMessage NVARCHAR(400);
    DECLARE @ErrorSeverity INT;
    DECLARE @ErrorState INT;

    SELECT @ErrorMessage = ERROR_MESSAGE();
    SELECT @ErrorSeverity = ERROR_SEVERITY();
    SELECT @ErrorState = ERROR_STATE();

    RAISERROR (@ErrorMessage, — Message text.
               @ErrorSeverity, — Severity.
               @ErrorState — State.
End Catch;

— Send back a small informational XML message to the Adapter.
— This XML message will be published in the Messagebox Database:

Select rowsprocessed,  ErrorCode,BatchIdentifier,
  (Select Distinct sch_prod_order_id  ProductionOrderId
   from #tempInsertNewProductionOrders  ProductionOrder
   For XML Auto ,Type )
from #results  Results
For xml auto, ROOT(‘BulkInsertResults’)

— Note with Sql 2005 can:
— a) Place a Root node around the returned XML
— b) Specify a Target Namespace.
— c) Do nesting with an XML Auto Statement.


A sample of the XML returned from this stored procedure is as below:

The above XML message is then passed back to the Adapter. This Small message is then published into the MessageBox Database.
Note: When authoring this stored procedure, any information can be returned in this XML message.

Once this small XML message is published into the MessageBox database, any number of orchestrations can subscribe to this message. In this particular case, the main message (Now residing in a table in a Sql Server database) is to be split into separate messages by using the distinct production orders returned back in the above
XML message. A discussion of this pattern can be found here. The productions order records in the Sql Server table could also be transformed (mapped) by using Select Statements with XML Auto and XML Explicit clauses.

Results of Tests

The following hardware was used: Laptop with 2.0 GHz processor.
Windows XP operating System hosting a Windows 2003 VPC image with 1.3 MB of memory allocated.
This VPC image hosted the BizTalk 2006 Server and a Sql Server 2005 instance hosting the BizTalk databases.
The local Sql Server instance also hosted the database where Bulk Insert operation occurred.

Test for processing one Message

Only one flat file message was processed at a time. Only one implementation was tested at a time:
Either the PipeLine (Implementation one) or the SqlBulkInsert Adapter (Impelementation Two).

Results for Pipeline Processing

Size of Flat File Time for Pipeline Processing

XML File Size (Published into the MessageBox)

Number of Rows (Production Orders)  in Message
201 MB 25 Minutes 767 MB 807428
108 MB 13 Minutes 413 MB 435076
56 MB 7 Minutes 213 MB 224130
30 MB 4 Minutes 114 MB 120234
15 MB 3 Minutes 56 MB 62234
8 MB 2 Minutes 31 MB 29123

Results for SqlBulkInsert Adapter

Size of Flat File Time For SqlBulkInsert Adapter Load Number of Rows (Production Orders)  in Message
201 MB 3 Minutes 807428
108 MB 1 Minute 15 Seconds 435076
56 MB 50 Seconds 224130
30 MB 15 Seconds 120234
15 MB 8 Seconds 62234
8 MB 5 Seconds 29123

Discussion of Results

Implementation One: Processing with a Flat File Schema and Custom Receive Pipeline

This is the usual method to process a flat file in BizTalk. Besides disassembling (converting) the flat file into an XML format, receive pipelines perform many other important operations such as:
a) Decoding
b) Disassembling
c) Validation
d) Custom operations using Custom pipeline Components. One example would be Unzipping a file.

The Good

a) It works!
b) From a design point of view, this is the preferred method as the pipeline can perform a number
of operations on the original message before the final message is published into the messagebox database.
c) Much more complicated flat files can be disassembled in a pipeline. The flat file used in the demo, is simple in structure.
An example of a more complicated delimited and positional flat file is as below, with Header, Detail and Footer Records:

US        Alice Smith         123 Maple Street    Mill Valley    CA 90952
US        Robert Smith        8 Oak Avenue        Old Town       PA 95819
Hurry, my lawn is going wild!

The Bad

For large Flat File Messages:
a) the operation is somewhat slow (see times above),
b) CPU is pinned the entire time the flat file is being processed in the pipeline.
c) The original flat file message when converted to a XML, more than triples in Size, for example:
201 MB (flat file) transforms to 767 MB (XML). Depending on the operation being performed, a 767 MB message may be unmanageable (for example as a source message in a map).

Implementation Two : Processing with the SqlBulkInsert Adapter

This implementation was discussed in detail above.

The Good

a) Much faster in pre-processing the large messages (see times above).
b) Will not pin the CPU on the BizTalk machine (pipeline processing is skipped)
c) Large XML messages are not published into the messagebox database.
d) There really is not much to this -> A custom adapter and a stored procedure.

The Bad

a) A more complex flat file (see complex flat file message above) cannot be processed with this implementation. This is a limitation of the Bulk Insert Statement.
Other techniques could be investigated to Load structurally more complex flat files into Sql Server including:
i) DTS (Data Transformation Services). Sql 2000 and Sql 2005
ii) Sql Server Integration Services. Sql 2005
b) For the initial message the receive pipeline processing is skipped. If the incoming flat file was encoded or zipped up then the incoming flat file could be first be processed by another receive location/pipeline to decode or unzip, and then routed to a folder where the receive location is configured to use the SqlBulkInsert adapter.
c) The process (orchestration) that subscribes to the small XML message published into the MessageBox, has to be configured to connect to the database where the main message is stored in a relational sql table. This might mean just configuring Solicit – Send Send ports using a Sql Adapter
d) If using a Bulk Insert Statement to load in the flat file, Sql 2005 might have to be used. (See comments in the Stored Procedure) 

Final Thoughts

I have only spent about two hours on the adapter and it is not ready for production (It is a prototype).
I have also tested by processing two large messages simultaneously. Therefore if you are interested in 
using it, download the code and modify or just start from scratch. The code in the adapter is relatively straightforward.

The adapter was originally created with the BizTalk Server 2004 Adapter Wizard and uses the common adapter base classes and interfaces.

This technique (Adapter) could also be used to Bulk load an XML message into Sql Server.
Download the Code HERE. Read the ReadMe.Txt before installing and running.






Wish List for BizTalk 2008 (Biztalk vNext)

With the official launch of Biztalk 2006 just weeks away, I could not help but start thinking about what I would like to see in Biztalk vNext (a.k.a Biztalk 2008 / 2009). 

It will most likely have support for workflows written using Windows Workflow Foundation or even better it may be built on top of the Windows Workflow Foundation. 

That is great and all, but it is important not to forget that Biztalk is an integration tool.  The support of workflows using Windows Workflow Foundation alone will not help solve some of the existing tricky integration scenarios we are facing today.

With that said, what improvements and enhancements would I like to see in the next version of Biztalk…

Convoy as a setting on the receive port

Many times convoys are used to achieve first-in first-out for various conditions.  This requires a Receive followed by a 2nd Receive and usually a timeout (using a Listen Shape).  Rather than using multiple Receives that add little value other than making a Convoy, add a setting on the Activate Receive Port so it can be marked as “FIFO” thus providing Convoy behavior based on the correlation set on that Receive Port.  This would be similar to the new FIFO support on the Send Ports in Biztalk 2006.

Ability to suppress event log messages via a context property

When using Delivery Notification it is common to have extensive exception handling built inside the Orchestration to react to any problems and possibly write to some other event log (other then Application).  It would be nice to be able to suppress the events written to the Event Log in this case.  Also, having an easy way to change what event log is written to would be nice.

Better control over property promotion

This applies mostly to direct message box binding with untyped messages – that is messages that are System.Xml.XmlDocument.  I am a big fan of direct binding to the message box.  This is an underlying key component to Biztalk Server even though it is hidden nicely to new developers.  It would be nice to have better control on what is promoted when sending messages out of an Orchestration.  There are some tricks using a correlation set to get values promoted that works.  But, I would like to see something a little better or at least to have this process documented. 


Checking a promoted property that does not exist does not throw an exception

It sure seems silly to have to put checks on promoted properties inside a Scope Shape just in case the value does not exist.   

When a default value is set on a promoted property and it is associated with a schema the default value should always exist in the message context

This can currently be done rather easily using a custom pipeline component but that can be a maintenance nightmare and error prone.  This would also help support a “Not Exists” subscription condition by allowing the default value to be treated as the “Not Exists” condition.  The value could be overridden in a pipeline for normal processing.  It would also help with the exceptions when checking properties that do no exist as noted above. 

Support for multiple Acks and NAcks with send port groups

It would be nice to be able to process multiple Acks and NAcks when using Delivery Notification inside an Orchestration.  This would allow for the use of send port groups for these scenarios.  I have been trying to build a process to do this but after MANY hours of work I have not been able to get it to work.

Feel free to post your own thoughts and comments.

Business Rules Enhancements in Biztalk Server 2006

I tend to see more and more clients using Rule Engines now then I did in the past.  So, what can we expect to see different and better in Biztalk 2006?

At the BPI Conference two weeks ago we found out some of the new enhanced features that the Business Rules Engine would have in Biztalk 2006. 


Some of these are:

– Ability to be installed without Biztalk

– No longer needs to be called inside an Atomic Scope

– Document type will try to be auto populated when you add a schema from an existing Biztalk project (this is great – until it gets it wrong)

– Ability to have debug information written to file at run time


Of course, these can change before the final release of the product early next year.


Business Rules Message Construction in Biztalk Server 2004

Wow.  I was in the middle of working on a blog post about message construction and the Business Rules Engine (BRE) in Biztalk 2004. 

Several other Biztalk Developers and I were talking about this topic at Tech Ed and at the BPI Conference.  I looked at Richard’s Blog and he posted an excellent post on the exact same topic! 

At first glance, it looks like the BRE changes an immutable message or does it?


The long and short of it is:


No: the BRE does not change an immutable message


Yes: the Call Rules Shape does construct a message without using a construct shape


Yes: you can lose Context properties in the are set on your original message inside the Orchestration (see Nishil’s post for more info)


That got me thinking… I wonder if you put a persistence point after setting the Context Property if you would still lose the value?


Twin Cities BizTalk User Group

On September 22nd we held the first Twin Cities BizTalk User Group at the Microsoft offices in Bloomington, MN.  This was a great opportunity to see and talk to fellow BizTalk enthusiasts and catch up with people I hadn’t seen in a while.

I would like to thank the other people who helped me put this user group meeting together:

   Andy Morrison,
   Scott Colestock,
   Jim Gaudette and
   Todd Van Nurden

I presented on BizTalk 2006 and the changes and additions that will be coming with the new version.  The presentation can be downloaded here.

If you are in the Twin Cities we would love to have you attend.  Our meetings are every other month and our next meeting will be Thursday, November17th, 2005 at the Microsoft offices in Bloomington

As far as we can tell this is the first BizTalk User Group in the country.  If there are others please post a comment – we would love to hear about it.