Running an Integration Competence Center for small businesses

A while ago I saw a post from an old colleague on LinkedIn where he linked to an old article called Too small for an Integration Competency Center? Think again. I was intrigued to read the article since I am in the middle of establishing an ICC at the organization I'm at, and we are quite small in that regard.

The four building blocks involved are:

• Adopt common integration practices
• Standardize on a common integration platform
• Create a decentralized group that specializes in integration
• Form a centralized ICC

This has all been done at the organization I am at, with the experience I have from larger organization as a basis for the setup. In my case, with an IT department of a whopping four people, forming an ICC is no big challenge in that regard.

I have opted to include external parties in the ICC to form a team that is not bound to a strict organizational hierarchy, but rather be agile and flexible to the demands from the organization at stake.

The basis of the ICC is in my role as integration architect/supervisor in the IT department. Coupled to this I have included an external integration architect to have a form of sounding board. Most, if not all, integration development work is carried out by consultants which the external integration architect is responsible to coordinate. As needed, I also include internal resources in the ICC to help forming the ICC and drive the integration development forward. So basically, the ICC is in it's simple form just two persons, with several others coming and going as needed.

So far, this is working very well. The agility in the ICC is a strong factor to consider. A lot of times, people are included full time even if they are not needed full time. Since the external architect is contracted on an as-needed-basis, costs can be brought down. Of course, that puts a lot of burden on me.

However, I have strived to create a simple and effective process of the entire integration life cycle. Document templates are available for everything from requesting an integration (done by the organization) to detailed specification and testing of the completed development. The development guidelines and deliverance specifications are to be as standardized as possible to not bind anything to a single person or organization (except our own). I aim to be able to both replace the consultants involved as well as myself without having to explain anything about the setup and behavior of the ICC and involved parts.

As for the four building blocks above, I feel that all (almost) of them can be ticked off in our case. The platform is BizTalk, best practices are established through documentation and processes based on experience from me and consultants, a decentralized group is formed and the ICC is formed and handling most of the integration needs. What is left is the funding that is still divided and handled by separate organizational units rather than taken care of completely by a central organization. This is currently the biggest issue to handle since it hamper a true value-for-money path regarding integration.

But to summarize, in accordance to the article linked in the beginning of this post, you cannot have a too small of an organization for an ICC to be formed. By including resources on a need-to basis as well as including external consultants where needed, it can be handled in a very effective way. The important part is to have a clear process of the entire life cycle as well as clear boundaries between different parties involved. My role in the whole ICC is most importantly to coordinate the different tasks between the resources, and as long as that is done appropriately, everything is flowing smoothly. Having a centralized funding of the work is not necessary, but very (very!) preferable.

Simple dos and don'ts for versioning webservice interfaces (and similar integration interfaces)

Earlier this week I went to a user group gathering for the ERP system we utilize at work. The topic for the day was "integration". At the end of the day during an open Q and A session, I approached the ERP representatives with questions on how they worked with the web service API exposed from the ERP system. I asked them if using the current web services guaranteed that they would work without changes and need for deep regression testing when upgrading the system. I also asked them how new functionality to existing services would be implemented, from an integration point of view.

The answer was that current implementations would require normal regression testing after upgrading, but that there could be changes in the backend making the services work a bit different in the new versions, depending on the changes made in the business logic and GUI.

This made me a bit quezy and fidgety. In the best of worlds, I want a web service to be seen as a hard written contract between two parties, in this case the ERP system on one side and the integration platform on the other. By exposing a service, you should guarantee that the specification for the service is not to be changed and that the underlying logic is to be set in stone as well. That seemed to not be the case.

The discussion then evolved into the realm of versioning and planned changes of existing services. As of now, new functionality is added to the service without changing its name, version or namespace. Fields are simply added as optional.

This method is something I have several issues with, and which I also expressed at the session (maybe a bit too harsh, but it is an important subject to me). First off, by adding new functionality as optional fields, you will break one of the (many) benefits with web service contracts: hard typing of them! By doing it this way, the developers will make future maintenance of the services harder, for both themselves as well as those utilizing the services. Also, the clients will have to upgrade the integrations whether they want to or not as soon as the ERP system is to be upgraded and the services change.

I instead proposed that they should version their services and operations. When changing an operation, create a new version of it and as long as it is possible, make sure that the previous version(s) are still functional. If a service operation is to be removed from the system, mark it as obsolete and let the clients know which version of the ERP system it will be removed in, and preferably how to utilize other services and operations to enable the needed functionality. I myself think that this is a very obvious way of handling the dilemma. I have grown to understand that it is very rare that versioning is used though and I do my best preaching the above statements.

So please, make sure that you use some form of versioning strategy when building your interfaces. It will make my (and many others, I guarantee) life a lot easier.

A take on documenting integration platforms using Visio

Earlier this year I wrote a post about having Visio update a field on a shape automatically when editing any other field, making it possible to always know when the metadata was last changed. In the comments I was asked to share the Visio sheet which it now is time to do. This is the result of that request, in some way documenting my attempt to use Visio to document an integration platform. Hopefully you can gain some ideas to use in your documentation endeavours.


In almost all of my encounters of different integration platforms at various companies, there has been a high level view of the entire integration setup showing the various message flows available. This map has usually been constructed in Visio. There has also been two types of layouts, lets call them the structured layout and the unstructured layout.

The following is an example of the unstructured layout. It's common and does a fairly good job at describing the full environment. It is often seen when there has not been a true approach to working with integrations.

Most of the time they are filled with objects, with varying descriptions, showing the entire network.You can see that two systems interact, but you do not know in which way. It is also often hard to see how everything is connected if you as in the example has an integration hub in the middle, doing some message routing.

Hence we get the structured layout as is seen below. In this case, there is a very solid structure with the integration platform in the middle, and the integrated systems on both sides with arrows showing each and every message flowing through.

These maps are good at showing the message flows as every message is entered, but it doesn't give you the overall picture that the unstructured layout gives you since a lot of fluff is stripped away. Usually there has also been an improvement from the unstructured layout in that each message flow has been given an identity which then maps to more detailed documentation in a file archive or document portal.


So, earlier this year I did a test with a variant of the structured layout, seeing if I could document the current state of integrations at my current job. The map I was given from the beginning was according to the unstructured layout. A lot of information was out of date, systems had been taken out, others taken in, versions changed, message flows altered and so on. I had to create an up to date picture of the integrations, so that I could commence working with them.

I took a brand new Visio document and started documenting. I created a "system" shape that had the following shape data:

• System name
• System name 2
• System version
• Server name
• Database
• Database version

In my Swedish Visio it looks like this:

In order to make it easier updating and changing these properties/metadata for all shapes, I changed the Shape Sheet to make the form data window open when doubleclicking on the shape. You can do this by editing the Shape Sheet property "EventDblClick" under "Events" to read "=DOCMD(1312)".

I then did the same thing for the arrow shape adding metadata that I want to know for each message:

• ID
• Object name
• Status (i.e. Planned, Under development, In use, To be removed)
• Transport (i.e. Web service, File, FTP, WCF...)
• Type (i.e. Xml, Flat file...)
• Frequency
• Automatic (i.e. a drop down giving the option of Yes or No, indicating whether the message transfer is completely automatic)
• Sync/Async
• One way/Request-response
• From system
• To system
• Comment

Many of the fields are defined in the form data editor as dropdown menus as to make the data more consistent and easier to edit. I also added the "From system" and "To system" in order to get good documentation automatically created by exporting the form data to Excel. As with the System shapes, I added the doubleclick event to the shape.

I also altered the shape so that the choices made in the form data will be visible in the shape. Depending on the "Status", the arrow will be yellow (planned), green (under development), black (in use) or red (to be removed). This is done by editing the Data Graphic and setting that each value will match a specific color.

Also, setting the "Automatic" to "No" will show a little man besides the ID and Object name. This can be set in the Shape Sheet as well under "Text fields" and "Value". In my case, the formula is: =IF(STRSAME(Prop.Automatic,"No",TRUE)," 웃","")&" "&Prop.ID&". "&Prop.Objectname&" ". But is of course dependent on the name you give each property in the form data.

In the same manner, I have a setting that will change the arrows on the shape so that I get an indication if it is a one way transfer or a request-response one, by adding a little unfilled arrow in the other direction of the data flow.

I have also adjusted where the text appears on the connector according to this post from the Visio Guy, so that the text is always at the end of the line instead of in the middle.

So, is this a good way of documenting message flows?

Well, in this case, it has been very helpful. I have a single document that shows me quite a lot of the current setup and state. It eases on the job of manually editing shapes to indicate whether it is a planned message flow, if it is request-response, a manual job and so on. Instead, I just select the desired property and let Visio format the shape accordingly, giving me a neat view of the state.

After documenting all message flows in this Visio document, I will most likely strip it down. I will remove a lot of the metadata in the connection shapes, and have this solely in the integration specification document that I will create for each message. I do not like to have duplicate information in my documentation, and I feel that that sort of data will belong better in the written documentation than in the Visio sheet. I will still keep the color coding depending on Status, as well as some other visual cues that is helpful when looking at the overall view of the integration platform.