Cloud-based Data Integration has been available for quite some time. But Cloud-based Application Integration (Cloud-based EAI) is a new arrival.
Vendors like SnapLogic and Boomi helped lead the charge into that Cloud-based Data Integration space, and heavyweights such as Informatica helped make it a mainstream technology.
The First Cloud-based Application Integration Platform
MuleSoft, a well-known vendor of open-source message bus software, is poised to change the Application Integration (EAI) landscape as well as the Cloud Integration landscape with its new iON Cloud-based Application Integration platform. MuleSoft's iON is the first full-function built-for-the Cloud Application Integration suite.
Data vs. Application Integration
Data Integration, which is a "batch mode" or scheduled operation, is about synchronizing data between different data sources and targets. It can involve a lot of data manipulation, reconciliation, de-duplication, cleansing, standardization and other data-intensive operations. It fundamentally provides a layer of abstraction from the underlying data sources. A Data Integration task might run once a week, once a day, once an hour, or a few times an hour, or even more often than that. It doesn't run hundreds or thousands of times an hour, though. A Data Integration task may involve a single record of data or Terabytes of data.
Application Integration is about the reliable and timely exchange of messages between applications (rather than data sources). It typically operates at near real-time speeds, and often includes workflow types of functionality on top of the core message-level guaranteed delivery capabilities. Application Integration is about enabling business processes that span multiple different systems, and provides a level of abstraction from the underlying applications and business processes. Typically the amount of data involved in a single Application Integration task is quite small (although there may be many thousands of Application Integration tasks that get processed in the course of an hour).
The following diagram (click for full-size image) shows a high-level depiction of the differences between the two different types of integration. Whereas Data Integration is about the movement and transformation of data from one place to another - ignoring application or business logic, Application Integration is about linking systems together at the application logic level - the data simply goes along as a spectator.
Deciding - Which Integration Method for You
As I've written before, neither method of integration is superior or inferior to the other - it simply depends on the need. For example, if you need to populate a large data warehouse from numerous transactional systems, Data Integration is probably the way to go. If you want to create a single business process that spans multiple independent applications, then Application Integration is the way to go. Nearly all enterprises with integration needs require BOTH types of integration technology.
Application Integration Example
As a way of explaining the benefits of Cloud-based Application Integration, it may be most effective to present it in the context of a common use case that most people can relate to - the "order to cash" cycle.
Order to Cash (OTC) is a common business process - every business that exchanges some sort of good for some sort of compensation has a version of this business process. Consider a scenario where a business sells physical goods to other businesses. They receive orders for goods from customers. They process orders, create invoices, ship orders, receive payments. This process spans potentially multiple different systems - which may span multiple organizations (as many companies outsource shipping). A high level diagram of this process might look similar to the following diagram (click for full-size image):
For a business process such as OTC, which is a single logical business process, Application Integration rather than Data Integration is the preferred approach, for a number of reasons:
1) Timing - Application Integration is event-driven and near realtime (rather than scheduled)
a. The faster the process completes, the faster the company gets paid - resulting in lower Days Sales Outstanding, better cashflow.
b. The faster and more accurately the orders ship, the higher the level of customer satisfaction
2) Integrity of the process is important. Application Integration frameworks guarantee that certain events will happen - and that they will happen in the correct order.
a. The order that the events occur in is important. For example, an order should not ship without an invoice being created. Orders always ship with the product and an invoice together.
b. As soon as a payment is received, financial data in multiple places needs to all updated "together" - customer records, corporate ledgers, etc.
Cloud-based Application Integration Will Change EAI
MuleSoft's iON marks the beginning of a tremendous and long-overdue shift in the Application Integration market, for at least three major reasons:
First, on-premises EAI platforms have been around for a long time - they've been mainstream since the late 1990's. They're highly capable, but have a reputation as being very complicated and expensive to implement.
Pretty much any enterprise-class software that ends up in the Cloud becomes tremendously easier to use and faster to implement. Compare Siebel on-premises vs. Salesforce's SaaS application for example.
It's not that the Cloud magically makes things better - it's that the software packages built for the Cloud are new, more cleanly architected, free of baggage and legacy problems. For example, many on-premises integration stacks have code dating back to the early 1990's, with user interfaces that would easily be at home in a Windows 95 environment.
They've been extended dozens of times and have become unwieldy, unproductive.
Secondly, increasingly more and more operational as well as analytical applications in the enterprise are moving to the Cloud. SaaS applications like NetSuite or Salesforce, and increasingly, Cloud-based analytics platforms are the default choice for those doing significant analysis and number crunching.
With the locus of applications and data moving increasingly towards the Cloud, that's where Integration increasingly belongs. Just as SaaS and Cloud are becoming not only mainstream - but preferred choices for business applications, Cloud-based Application Integration will become the preferred choice for EAI.
Finally, MuleSoft's iON follows a SaaS pricing model. Integration products like TIBCO can easily cost hundreds of thousands of dollars - typically payable up-front, with yearly maintenance easily costing tens of thousands of dollars. MuleSoft's iON has multiple tiers of pricing ranging from $150 a month to $1,000 a month and a pay-as-you-go model.
Much like Salesforce with its $25 per month entry-level cost upset the Siebel Systems (whose competing product was often hundreds of thousands of dollars) applecart, MuleSoft's iON will do the same to the EAI marketplace.