4 Steps to Consider in a Cloud Integration Framework

By Guest Author  |  Posted 05-11-2016 Print Email

The transition from only a handful of monolithic systems to a more disparate ecosystem of applications has made cloud integrations more challenging.

By Matt Johnson

How do your organization’s critical systems talk to one another? What does your integration landscape look like?

Traditionally, organizations have built ad hoc, point-to-point connections, but this approach creates a tangled web that requires extensive maintenance and often leads to performance lags. And, as the shift to cloud computing becomes even more mainstream, enterprises are now realizing that these traditional approaches to integration do not scale effectively or provide the capabilities needed to properly protect information that resides outside of the corporate firewall.

Furthermore, the transition from having only a handful of monolithic, enterprisewide systems to a more disparate ecosystem of applications has made these integrations more challenging and added to the need for a clear strategy.

In light of these changes, now is the time to develop a comprehensive and forward-thinking integration strategy. Doing so will allow you to take full advantage of the agility of cloud-based solutions while continuing to use existing on premise infrastructure investments.

The following four steps will provide a framework to help you develop an integration strategy that’s appropriate for the cloud era.

Step 1: Look Ahead

First and foremost, you need to assess any potential changes ahead and the direction of your five-year roadmap.

From a business standpoint, you need to have a clear understanding of your objectives and priorities, including any merger-and-acquisition plans. Understanding these business priorities at a high level and ensuring alignment will help define your integration strategy.

Meanwhile, from a technology standpoint, you need to consider how game-changing technology such as cloud and mobile will impact your organization. For example, how will you secure, scale and manage a cloud and mobile environment that stores data outside of your firewall and provides users with real-time access to information? Will you introduce any Internet-of-things technology? How about replacing custom or in-house solutions with SaaS solutions or moving to a micro-service-based architecture?

In general, the key here is to recognize that you need to stick your head above the water in order to plan your strategy.

Step 2: Understand Your Current Data Landscape

Once you identify where you’re going, you can come back to present day to see where you are currently and how you will get to that future state. Start by looking at the complete lifecycle of data in your enterprise: Is it transactional? Is it structured? How is it staged, moved, replicated, accessed and archived? You’ll likely find that many past integrations were batch/schedule-oriented, occurred behind the scenes and used a patchwork of different legacy technologies.

It’s often helpful to start this process by identifying each of your technology stacks. A stack is a layered approach to viewing your technology in which each enterprise competency domain (e.g., customer engagement, administrative management, performance management, request servicing and any core competencies) is automated by platforms such as CRM, ERP, workplace productivity, service management and homegrown systems that have their own stacks of integrated applications. You can read more about this approach here.

Although this can be a daunting task—in fact, it is typically one of the hardest parts of developing an integration strategy—it will reinforce your activities to modularize integration efforts down the road and, therefore, the payoff is well worth the effort.

Step 3: Develop Your Strategy

In the past, system integrations have been treated as projects, meaning they have had clear start and end dates. However, it’s time to rethink that model. Things are only getting more complicated as the number of data sources and integration points grow, and, as a result, integration should become its own core competency. By this, I mean that it shouldn’t just be part of a project, it should be a full-fledged strategy that focuses on people, tools and processes. Having a dedicated team revisit integrations regularly is also important because of how often solutions and APIs change in the cloud.

As you develop your strategy, be sure to take business processes into account (you should never create an integration without an explicit business case in mind), establish integration principles and patterns and create an enterprise integration playbook. You can use the data lifecycle use cases you previously identified to develop a standard approach to common use cases that you can include in this playbook. This playbook should also help with governance, though you may want to take governance a step further and add a center of excellence to oversee your strategy and related activities.

Additionally, you want to resist the urge to build integrations yourself. Instead, you should standardize on an enterprise-class technology integration solution (and you should keep the number of these tools in play to a minimum in order to reduce complexity). Doing so will help increase agility and prevent you from creating a tangled mess of integrations.

Throughout all of these efforts, it’s essential to remember that the cloud increases the level of complexity for integrations since data will move outside of your firewall and APIs are not under your control, so you need to be more sophisticated in your approach.

Step 4: Apply and Iterate

After you’ve developed your strategy, it’s time to set it in motion. As you do so, make sure that everyone understands its goals and direction. One approach that I’ve seen work well here is promoting it among internal teams as part of a service catalog so that it can be managed and identified along with current solutions. When done correctly, this approach can help improve standardization and simplify management.

Most importantly, remember that integration is not a one-time event. Rather, it’s a process. You may not get everything perfect the first time around, but continually going through your integrations and asking questions such as what they are doing, how they are working and how they are changing is essential to effective integration management.

Matt Johnson is vice president of advisory services at Cloud Sherpas.


Submit a Comment

Loading Comments...