As you’d expect some interesting insight from Gartner here. I’ll leave you to digest the report, but wanted to pull out the key findings, recommendations and one of their missed 2010 predictions for some commentary:

Key Findings

Over the next few years:

  • The cost of doing application integration will increase as companies do more B2B e-commerce, information processing, and mobile, social and cloud computing.
  • The accommodation of more mobile devices, alone, will substantially increase the amount of work — and cost — required for application integration.
  • Companies will soon do more B2B integration than application-to-application (A2A) integration.
  • Companies will spend more on application integration than on new application systems.

I can certainly see even on some smaller businesses that the need to integrate with 3rd parties is growing. Especially information businesses. Companies are either providing feeds out or consuming feeds, but on the larger enterprises there’s a raft of supplier related integration as well. The findings also cover the extension of internal applications into the cloud. Cloud applications are now an integral part most IT operations even if they aren’t owned within the IT organisation, which is fairly common. The need to connect these systems into the on premise environment is the challenge that lands at the CIOs door whether or not they own the application.

That would logically drive this increase in application integration spend. In a lot of scenarios, on business already engaged in integration, the requirements to increase spend on tools may not be required. Look again at your suppliers, internal spend and renewals. There are a lot of integration customers out there sat on expensive solutions or home grown integration that is difficult to quantify the cost of. Jitterbit have been able to help businesses reduce their spend on integration by swapping out more expensive solutions. The barrier for most businesses on doing this is the perception of complexity involved in migrating existing integration solutions.

The hard work on integration is normally done in the initial project delivery; understanding the end points, business rules the drive the transaction . Actually changing integration product is less complex than people anticipate. We had a large customer in the UK recently migrate their Salesforce integration across to Jitterbit with little or no input from us.

Recommendations

CIOs and other IT leaders should:

  • Plan to increase IT budgets for application integration over the next five years.
  • Include integration as a core competency to help enable B2B projects and all major IT initiatives, including the Nexus of Forces (mobile, social, cloud and information).
  • Develop an overall integration strategy, including an integration competency center (ICC).

Look first at what you existing integration is costing you. IT departments often get bogged down in Service Oriented Architecture nirvana over and above delivering business value. Ultimately the business cares less about how you deliver and more about how fast/cheap you deliver it. That might be in some enterprise eyes a small world view of the integration challenge but the reality is Integration architecture projects are often like old data warehousing initiatives. They are unnecessarily expensive and take too long to deliver.

Strategic Planning Assumption: By 2017, over two-thirds of all new integration flows will extend outside the enterprise firewall.

This means your integration platform needs to be support both SOAP and REST with considerably more effort going into JSON based REST APIs. That doesn’t in itself rest that well with the SOA route where SOAP has ruled. I’m not a huge fan of REST APIs in that they are more often than not poorly architected. They are typically built for mashup over and above enterprise integration requirements. We need to be able to do it though and cannot knock SaaS applications off the integration radar because it doesn’t fit our architecture view. Fortunately REST is well within Jitterbit’s crosshairs with XML already supported and JSON available in the upcoming version 5 release.

These APIs have more often than not been built by web applications developers for web application developers and moving forwards we’re going to need our suppliers to build their APIs for the enterprise to engage.

Missed: 2010 Prediction — By 2014, some form of packaged IaaS will be used to address more than 75% of all cloud-to-cloud service integration requirements.

My view on this is that organisations aren’t using IaaS or iPaaS because it doesn’t work for them yet. The solutions aren’t powerful enough or don’t have sufficient traction. Connectors may be part of the reason, but the majority of connectors are built for specific use cases. They fit a large proportion of a vendor’s base, but they don’t handle complexity. Organisations seem to court business process complexity whether or not it is necessary and that often causes pre-built connectors to fail. It also creates this barrier for IaaS or iPaaS. Some business processes just don’t fit the use case connectors were built for.

Control is another reason that on premise integration remains first choice although increasingly customers are looking for a one stop shop solution. This should perfectly place vendors of iPaaS solutions, but they just don’t appear to be there yet. If you’re reading this then it’s likely that integration is firmly on your radar. Contact us to discuss you integration challenges for 2013.