Cloud integration is no longer a point of contention when selecting a vendor;with countless best-of-breed SaaS solutions available today,business and IT professionals alike accept that connectivity is a critical box on the cloud application checklist. In the frequent scenario where thevendor does not supply the “out-of-the-box” solution you need, the good news is there is a proliferation of 3rd party tooling that supports many levels of integration to support your business solution. But this forces questions: what do you have to look for in your integration approach? What matters most to your business, and how can you avoid three common integration pitfalls to drive long-term success?
Integration Pitfall #1: Forgetting to ask the rightquestions
Some cloud integration pitfalls are easy to spotand avoid by askingdesign-oriented questions like: can the integration approach – using either 3rd party or in-house technology– support the volume of data I am moving? Can it transform the data in the manner I need it to? These are the questions most integration vendors answer when pushing out to market, proving that they are “best in class” for all the feature functions that IT typically looks at. Neglecting to ask these questions sets you up for failure, but you also need to look further than the design phase of cloud integration.
Integration Pitfall #2: Focusing only on short-term goals
When looking at cloudapplications in the long run, theinitial design phase only takes up a tiny percentage of time and resources compared to the lifetime effort of running that integration, so isn’t it more important to know how to support the integration going forward? Instead of succumbing to the pitfall of focusing entirely on design-side issues, IT would be better served asking questionsto support a cloud integration lifecycle management approach, including: will the integration keep up with the application on the other end of the sync? Will it let the right people know when something breaks, and provide an easy manner to correct issues?
The capabilities that a cloud integration approach truly needs to be “best in class” are those that help it keep data moving between endpoints for the entire integration lifecycle of design, develop, deploy and run. That means that the best cloud integration approaches must be able to:
- Maintain parity with each endpoint application
- Support notification of error conditions
- Facilitate creating great documentation
- Allow for easy correction of errors that do occur
- Provide support for changes in deployment models (cloud/premise/hybrid)
- Allow for evolution of the integration as use of the business application changes
These features focus on the cloud integration stage most-often overlooked, yet perhaps most critical – the run stage. It’s great to have the very best designer UI for a tool, or to offer the most complete catalogue of end-to-end solutions, but without the focus on keeping an integration running healthily and happily for the long term, these are just window dressings. The integration approach you select is only as good as the stability of the integration you build with it. The more fragilethe integration is, the more fragile the product perception is, and the easier it becomes to re-platform. This is not to say that having a great user experience is unimportant; it simply must be a cog in the bigger picture of the integration lifecycle.
Integration Pitfall #3: Building a static integration
Your cloud integration approach – whether it relies on a graphical integration interface or another tool – needs to be quick to pick up, deep in its capabilities and it must facilitate a faster deployment process than standard custom coded integrations could. But, theapproach also needs to avoid the pitfall of getting stuck in a given stage of the integration lifecycle – it must be flexible enough to flip between lifecycle stages, so thatsomeone else can come in and pick up where you left off, or switch from run back to develop for new capabilities. Over the course of your cloud application’s lifecycle, your business needs will likely change,and therefore the integration must change with it. If you do not have flexible tooling that allows you to not only understand what the original integration is doing, but to also easily make changes to the flow of data, it again can create a negative perception of all tools involved. The better the integration, the easier it is to manage and evolve, and ultimately the better the perception and rate of adoption for the core business applications involved.
Recognizing integration as a key value driver
Integration is a funny thing; in a perfect world, no one knows about it, no one talks about it, but everyone reaps the benefits of it. Integration has the Midas touch in these scenarios.Done correctly it can provide incredible upsides to the entire project, but, if done poorly, it can herald the destruction of a project just as quickly. When you enter an application project, make sure you think globally about the solution you are providing. Making an integration work is not just a task within a larger project, and it takes ample consideration and proper tooling to ensure that the consumers of this project never have to think about integration.
Allowing the consumers of your integration to easily become the caretakers of it, without a huge upfront investment of time and energy, will allow it to function at its peak for years to come. To get there, you need to start with the right integration approach that accounts for all phases of the integration and application lifecycle – not just the design and develop phases, but also deploy and run – and allows you to quickly transition the integration between teams and lifecycle phases without slowing down or breaking. By planning for this approach with the right tools and mentality from the outset, IT can ensure long-term success for both the applications and their supporting integrations.
Brendan Peterson, Product Marketing Manager, Scribe Software