Don, Brian, Great write up, and it confirms what we've seen and are worried about. You mention that you are planning to develop your own integration framework. Any hints as to what your approach will be? (without giving away the secret sauce :) Also, do you think that this situation will make the integration tools that (for instance) 2Increase had/has relevant again? Rob vanBrandenburg

Rob, We have had our own integration framework for over 3 years now. We got tired of rewriting the front end of our integrations with each succeeding version. As you know, AIF had a lot of downside and we felt it would be more effective to develop a true integration framework. Then comes along AX7. We sincerely tried to follow the Microsoft lead on the backend of our product. As you know, we stay as close to standard as possible. There just simply was not enough there to put together a consistent outcome. It is my personal belief, that unless Microsoft takes a different direction, this is going to hurt a lot of people. If people like us with 10+ years of AX development experience find this incomplete, what is it going to mean to all who are trying to follow the Microsoft lead. When we developed our integration framework EIF (enHanced integration framework), it was primarily to reduce our cost and time to produce and maintain integrated solutions. We have other partners that are now using it. We made it modular to enable people like yourself to simply plug in additional elements. Regarding 2increase, I will be talking to them at AXUG Summit. We are trying to see if there are common areas of interest in this area. If you are going to be there, I would invite you to join us. Although we have solved the problem for ourselves, I am hoping Microsoft will address this issue at the Summit. They need to be more open to the problems they are currently having with AX7. Don Riggs

The article covers many different things, so let me focus just on the main topic of AIF, DIXF and web services. Your description of web services seems to talk about custom services, ignoring the fact that they can be used for other purpose (web services are basically just a communication protocol). Most importantly, AX 7 exposes data entities as OData web services (not mentioned at all in your article), in a similar way to how AX 2012 is able to expose document services as SOAP web services. Your concerns about web services don't apply in this context. It's understandable that you miss this feature of document services, if you're not aware of OData services in AX 7. I also can't agree with the statement about DIXF saying that "No matter how complex the data set, you could only import one table at a time." On the contrary, the main purpose of DIXF is getting rid of working with individual tables, and using the higher-level concept of entities instead. That's what AX 7 continues to do with data entities. Martin Dráb

I've personally worked with AX for 15+ years and seen all changes come and go. Manually written file imports/exports, obsolete 3rd party products, COM/Business connector, Commerce Gateway, Biztalk adapter, AIF. new default services in 2012. I'm no fan of AIF, which is bulky, slow, tricky to modify and hellish to debug. But is has been there since 4.0 and the basic EDI-functions have been there out-of-the-box. Sending PO, receiving SO, sending and receiving ASN:s at least. If MS decides to throw it all away in it's new version, that's fine, but they should provide the same functionality and guidelines how to move up. I agree that Data Entities doesn't cut it. You give several reasons, but first and foremost is that you no longer can trigger an outoing EDI from the AX interface, or when posting journals, and there is no message logging at all, if using custom web services. And Martin Grab. I think you are missing the point. OData is a fancy and modern technology. However, in the real world I meet as a consultant, companies are stuck with old file based integrations and almost always an EDI-infrastructure. I see that there at the moment is a significant dip in willingness to invest in AX 7 and no interest in upgrades whatsoever. The reason for this is the limitations of Azure and the fact that all integrations have to be rewritten. In fact, we are not even sure how to rewrite some of them. IoT hub or Logic Apps may help solving some issues, but they aren't even on the market yet. Henrik

Martin, Thank you for your input. We are aware of OData and the differences to traditional web services. This article was not aimed at the technical differences in web service offerings. We did mention web services because they are part of the Microsoft messaging strategy. At best they are an enabling technology not a solution. Regarding the DIXF comment, if you put it in a historical perspective, DIXF was released with no capability to treat individual tables as a complete dataset. This ability occurred at some point later. That is why I said that it was my understanding it was solved. Part of the reason I wrote this article was to alert people to the fact that there are a lot of unanswered questions about integration. My hope is it actually get comments from people like yourself. If you or anyone feels that the Microsoft solution does represent an effective strategy, please comment. You can also submit your own article. We are searching for answers through this process.

Henrik, Thank you for you comments and the benefits of your experience! I to have been working with AX for 13+ years. Having worked with and helping design ERP systems, I was shocked at how little thought and efforts Microsoft has put into their integration layer. I am hoping we can get their attention.

Thanks for a great article Don. I think now that more information is available on Logic Apps that it's pretty obvious that Logic Apps will be the mechanism for building AX integrations, maybe using Flow for the simplest cases (but they'll have to be very simple for Flow). I'm not sure when all of the AX connectors will be available but I see Edifact and X12 mentioned as Logic Apps connectors (https://azure.microsoft.com/en-us/documentation/articles/apis-list/). Sadly I don't have an AX 7 project on my desk just now, but interesting times are just around the corner

Tim, Thank you for your advice. I am taking a great deal of heat regarding this article. The purpose all along was to start a dialog that will lead a solution. Yours is the first information I have received that has promise. Do you know if this is an architecture wherein additional connectors can be developed? Could we have further dialog?

Hi Don, I found your article very informative and refreshing. More of it, please :) It is important to point out the utter foolishness of removing working (although not dernier cri) tools to replace them with poorly thought out, one size fits all, "new" "tools" that are not up to the task. It seems to be the bane of our times, and not only on AX, or NAV, meanwhile almost everywhere where too many ideas only have to stand the test of a powerpoint presentation to people that don't know what they're talking about. It looks like this from the outside. Instead of being angry, these people should get to work to provide their partners with usable tools. with best regards Jens