Those who build cloud-based systems, such as myself, understand the strategy needed for data integration. However, many who work to bring cloud computing to their enterprise don’t put the same value on this technology.
I was very interested to see that Network World did a good job covering this important topic. This article by Rob Fox reinforces what we’ve talked about here for years: Ignore data integration, and you’ll have half a cloud computing solution.
As Fox put it:
“Although the cloud is becoming an increasingly ubiquitous delivery model for organizations of all sizes, IT leaders and other stakeholders should remember that the cloud isn’t an all-in-one solution. No single cloud services provider can fulfill all of an enterprise’s various requirements. In the majority of cases, businesses making the move to the cloud will require the services of an array of providers, combined with traditional on-premise application-to-application (A2A) and business-to-business (B2B) systems. As such, there is an increasing need to adopt integration strategies that support a multitude of complex integrations: A2A, B2B, on-premise enterprise applications to SaaS/cloud applications, and cloud-to-cloud (C2C).”
The reality is that most cloud computing solutions are not monolithic. They typically require what’s called a multi-cloud solution. This means that many cloud computing systems, sometimes private and public, are culled together to get to the solution the enterprise requires. For example, you might have database-as-a-service from one provider, IaaS from two others, and a PaaS from a third.
The idea is to mix-and-match these cloud solutions to form an integrated system that can share both processes and data in real time. Of course, multi-cloud systems are widely distributed, and need to be well planned. This includes operations, performance, security, governance, and, most importantly, data integration, that allows information to flow between the cloud-based components and other external enterprise systems.
These days, most cloud computing-related data integration is focused on the simple process of replicating data from a public cloud to an on-premise system. This typically uses data integration technology as a way to synchronize enterprise data between traditional on-premise systems, to those who have relocated to the cloud. For instance, sales data resides in a SaaS-delivered CRM, and perhaps delivery systems that reside within the enterprise data center. The data integration software works to make sure that the sales data, and other information, matches in both systems.
This data consistency needs to exist for any cloud-based system to work. Moreover, the data integration software should be able to communicate with any number of cloud-based systems (IaaS, SaaS, and PaaS), as well as traditional enterprise systems, including databases and enterprise software.
As Fox states, “When it comes to cloud integration, there are two fundamental questions that should be asked. The first is: What problem does the integration solve? Cloud integration refers not only to integration between different cloud-based systems, such as Magento, NetSuite, and Salesforce.com, but to integration between cloud-based and on-premise systems. Therefore, it is crucial to first determine what exactly is being integrated and for what purpose. In many cases, enterprises will need to accomplish both cloud-to-cloud and cloud-to-on-premise integration, but understanding the goal of any one integration project is a baseline requirement.”
As Fox points out, not all problem domains are “cloud-to-on-premise.” Indeed, there is a rising need to support cloud-to-cloud, certainly when considering multi-cloud-type deployments. This is a much more difficult issue to solve, considering that you’re dealing with two or more systems that reside outside of your firewall, and thus outside of your control.
The core nugget of advice that I received from the article, and advice that I dispense as much as I can, is this: When moving to cloud computing, integration must be thought about every step of the way. Along with security and governance for that matter. As Fox says, “Avoiding the pitfalls of this siloed approach to integration projects is critical, and businesses must build out an explicit integration sourcing strategy as early as possible in the cloud adoption process.”
This means you need to design and plan for data integration before the cloud migration or cloud development projects begin, when the cloud systems are implemented, and certainly during operations. Missing these steps results in a final cloud-based system that does not meet the expectations of the business. You’ll have information that is either not shared, or shared poorly within the final implemented cloud-based system.
Data integration is the single most important thing to consider when designing, building, and implementing cloud-based systems. As we get deeper into the use of cloud-based resources for enterprise IT, the need for data integration technology leveraged within these systems is an imperative. It’s something I’ve been saying for years, and I’m glad to hear others say it as well.