Why is Cloud Integration Still an Adoption Barrier?


One of the more puzzling topics in enterprise cloud these days is the market gap that seems to exist around cloud integration.  In survey after survey, integration is named as one of the top barriers to cloud adoption in the enterprise (see here, here and here).

The strange thing is that there’s no shortage of cloud integration offerings in the market today.  In fact the cloud integration market now also has its own acronym iPaaS, for integration Platform-as-a-Service.  Cloud integration tends to be focused on applications and data, and comes in a couple of flavors:

  • Cloud-to-Cloud — integration between two cloud-based applications or services and associated data.
  • Cloud-to-Ground — integration of SaaS and cloud-based applications to legacy, on-premise applications and data that resides behind the firewall.
Offerings are based on a varierty of technologies from legacy ESB (enterprise service bus) platforms to services built on Web 2.0 technologies and RESTful architectures.  Major categories of vendors and solutions include:

So if a plethora of options exist and integration is still an issue, what’s the problem?  A couple of issues we’ve heard anecdotally that could be in play here include:

  • Scalability — one issue we’ve heard repeatedly is that current solutions don’t provide required performance at enterprise scale.  Solutions that may have been initially designed with the SMB market in mind may be architecturally constrained, and not robust enough for the enterprise market.
  • Schema — one of the fundamental challenges with integration is always the mapping of two different data schemas between source and target.  While most iPaaS services provide a great deal of flexibility, we’ve heard that issues still remains around underlying data models.  A disconnect may exist between how SaaS vendors have defined their standard data models for supported business processes and workflows, and the highly customized data models enterprises may have designed in conjunction with custom, legacy on-premise app deployments.

We’d be very interested in hearing the practitioner point of view on the topic.  Is data integration holding your organization back from further SaaS or cloud adoption?  Why?  Where are current solutions falling short?

Comments

  1. A good article, I think personally that there are a combination of reasons holding back the cloud integration world.

    Data integration is important, since in the cloud world where API’s can change rapidly with new versions you need to be able to react a lot quicker. Therefore the profile of the integration projects is different. This leads to increased cost, and the need to have integration expertise on hand – in effect changing the role of the in house IT resource.

    I think that scalablity is less of an issue, its not a problem that I would say is different to the cloud or on-premise integration. It just tends to be more constrained in back office integration vs in the user world.

    You have the much wider issue of SLA and management of the integration layer, along with the question about where do you host the integration – logically that is something that itself belongs in the cloud. This then opens up the option of integrating to on premise applications and between cloud applications.

    Im not sure I agree with ESB’s being legacy, they are just adapting to a different use case – having a single point of control – whether it is in the cloud or on-premise is a big step to helping companies win the integration challenges that you talk about.

    The key in my view is having an integration layer that can be deployed anywhere, and can grow and scale with the organisations needs. This is where I think vendors are still finding their feet, since integration needs to be priced differently to allow this new world or as much or as little integration as you need – deployed anywhere.

    • Hi Jeff,
      Thanks for your comment – you raise some interesting and valid points. The need for a common integration platform that can be deployed both in the cloud and behind the firewall is an interesting one. It’d be great to hear which cloud integration use cases organizations are preferring to address via cloud vs on-prem models…
      Scott

Speak Your Mind

*