The first day of Integrate 2018 was exciting. Key speaker John Fancy started off with a walkthrough of IT from the 80s to the present. He talked about the big changes and how fast they have been implemented and effected IT in the recent decades. He made the comparison as to how slow things were changing in the past and gave an example about transferring money. It took the process of transferring funds hundreds of years to evolve from physically moving funds from location A to B, and now its just a few clicks away over the internet.
The point of Fancys introduction was that change is coming whether you like it or not, and you better be prepared.
He talked a bit about three stages you go through while dealing with change:
The first stage is denial and its basically you turning back on change, thinking its not necessary to adapt or evolve. The second stage is when you start asking the right questions and being curious about change. The third stage is when you see the big picture and know what road you have to take.
He quoted Tina Fey, “Say yes and you’ll figure it out afterwards”. And by that he meant that if you say no to change, someone else will say yes and leave you behind. If you say no, do you expect them to ask you the next time something comes up? So according to Fey and Fancy its better to say yes and to figure it out as you go.
Kevin Lam and Derek Li held the first product presentation which was an introduction to Logic Apps. They had some basic demos about how easy it is to get started. Then they had a slightly more advanced demo demonstrating the Azure Vision API to interpret a scanned invoice. In the end, the amount of the invoice could then be calculated.
Some announcements were made, such as:
It was a good presentation but the session was more targeted to people that haven’t used Logic Apps before.
He had some good points on where Functions fits into the introduction landscape.
He also had an interesting demo where he sent 30,000 requests against a Function to show how it automatically scale up. He also gave some tips on how to control how the scaling occurs. He was a bit short of time at the end, but he managed to introduce the concept of Durable Functions. Durable Functions is a way to write stateful Functions, which can be used to implement workflows.
This was actually the first presentation where BizTalk was mentioned. Paul Larsen and Valerie Robb announced Feature Pack 3 to BizTalk 2016. They focused a lot of new Office 365 adapters to handle contacts, mail and calendars. They also spoke a part of compliance with GDPR and FIPS privacy standards.
Compliance with GDPR and FIPS will also come in Cumulative Update 5 for BizTalk along with some other awaited features.
Miao Jiang talked about "The rise of APIs" and why we need an API Management platform to manage our APIs. He introduced some basic concepts in API Management such as policies, named values, products and subscriptions.
He made a few announcements such as better Application Insights integration, more metrics and KeyVault integration.
Clemens Vasters held this presentation and he had a little different focus than the previous speakers. This talk was more focused on architecture at a higher level and a little less on focus on specific platforms.
He made it clear that services should not be scoped on how many artifacts or how much code it contains but ownership.
Messaging: Where the publisher has an intent or expectation on what the consumer should do with the message. This would typically be something that would be implemented with Service Bus.
Eventing: Informs consumers that something has happened and don’t care much what the consumer does with the information.
He also defined to types of events
Discrete events Independent and immediately actionable. For this Event Grid would typically be used.
Event series Typically for streaming purposes where thresholds are monitored.
To start of his session Dan showed us the ever-growing amount of information that is being processed through the cloud environment. To handle this Dan gave his findings and experience on how to think when choosing the appropriate data handler for different types of data. Dan’s primart focus during his session was the Azure Event Grid and the Azure Event Hub. To make a gross oversimplification
Dan started of talking about Azure Event Grid which is a platform for consuming data from many different providers. He introduced a couple of concepts to bear in mind when working with Event Grid. Communication between applications/organizations, individual message (meaning that the messages don’t depend on one another), push semantics (trigger event), pay as you go, fan out. Dan also mentioned new features like being able to limited the amount of retires, hybrid endpoints, a dead letter endpoint for unconsumed messages/events.
Azure Event Hub was the next topic of the session. A big part of this part was dedicated to Kafka, which is a new Event Hub which is open source. Dan provided us with a demo showing the ability to consume messages from the Kafka Endpoint and receive these in the Azure Event Hub. The Azure Event Hub is used for fan-in purposes, acquiring big amount of messages in a stream(think of it as a ordered sequence), strict ordering.
To conclude this session; Dan shares his point of view that Azure Event Grid should be used for fan-in of data, where Azure Event Hub is meant to be used for Fan-Out Purposes and leaves us with this thought; The best tool depends on the context.
Divya Swarnakar and Jon Fancey entered the stage with a session about Enterprise Integration with Logic Apps and updated the participants about the improvements made during the past year, and things to come!
If you recall from the previous session where they announced the SAP connector for Logic Apps now being in private preview a demo showing the new connector and it’s functionality. As we all now integration between SAP and BizTalk from our previous experience working with these systems the connector also provides specification possibilities within the connector regarding namespaces. Based on the namespace specified the connector will listen for the specified namespace and trigger once a message containing the appropriate namespace is provided. If the namespace isn’t specified the connector will consume all messages sending it to a specific SAP ProgramId.
They also provided us with additional functionality that they are currently are working on for the SAP connector.
They also demonstrated a lot of new features that they are currently working on for enterprise connectors
Jon Fancy they took the stage and talked about improvements made on the mapping plain of the integration account. They have added support for XSLT 3.0, and briefly showed liquid templates, this was announced last year so consider this a quick recap.
A demo of the OMS template for Logic Apps, they also showed bulk resubmit feature and tracked properties. The tracked properties can now be added through the designer and not only the codeview. A coming feature is also the bulk download feature and an identification method for resubmitted runs of a Logic App.
And last but not least they gave us some information about things to come for Enterprise Integration. The biggest announcement is that they are currently working on a CONSUMPTION based pricing model for the Integration Account!
The last pass for the day was Kent Weare - Prinicipal Program Manager Microsoft Flow's presentation about Microsoft Flow. Microsoft flow is a "simplification" of Logic Apps. The difference between Microsoft Flow and Logic Apps is that Microsoft Flow does not provide any code view to implement or inspect its components. Office 365 users have access to this so be sure to expire, this is all about CS, AIM, FAS. Then, Kent talks about BAP (Business Application Productivity). In the middle of BAP, we find PowerApps & Power BI, which is targeted at power users. Flow and CDS (Common Data Services) are used to lock data in and out and / or apply business rules along the way before the information is sent to the subscriber of the information.
Flow enables application development / flow development for end users of the product, meaning that you do not have to have been taught in Machine Learning to be able to use its functionality. An example of this is Microsoft Flow's connector against LUIS (Language Understanding Intelligent Service), this was also demonstrated in a demo for a power cutter demo. This enables less involvement from IT departments to be required and, therefore, has less impact on critical IT processes that the IT department is potentially working with.
During Kent's session, four demonstrations of Microsoft Flow applications were implemented.\
Last but not least, a roadmap for Microsoft Flow was presented