Hosted applications remove the burden of buying and maintaining hardware, but off-premise systems could cause data integration problems that require serious consideration in any purchase decision.
Through industry standards, application programming interfaces (APIs), and middleware for enterprise integration, manufacturers have made great progress in uniting the islands of information that once kept them from sharing data between plant floor and enterprise systems.
As a result of these enterprise integration efforts, companies can now connect production operations with back-end accounting systems to keep material costs under control and manufacturing schedules on track. They can integrate supply chain data, warehouse management systems, and even sales forecasting applications to deliver the right product, to the right people, at the right time. When done right, it’s a huge competitive advantage. But any CIO or systems integrator will tell you that integrating applications that target different departments is not easy to do, since business needs change as fast as applications are upgraded.
Enter software as a service (SaaS) applications, which might seem to be an IT manager’s dream: no server and storage systems to buy and maintain. But their emergence presents a whole new integration problem between on-premise legacy apps and those that live in the cloud.
According to the recent InformationWeek Analytics 2011 Enterprise Applications Survey, 43% of SaaS users are very happy with the ability to deploy the applications quickly, but are much less satisfied with the complexity of integrating hosted apps with on-premise systems and data sources.
The InformationWeek article cites a handful of SaaS-based problem areas that are causing many CIOs to forgo the cloud, for now. Among them was the view that a lack of integration creates information silos—catapulting us back to the islands of information that make it difficult to share information or run business analytics. Another issue respondents cited was the inability to maintain a master data set that includes enterprise wide governance of the information flowing among the applications.
One has to wonder whether some of the new integration appliances and brokering services emerging, including Jitterbit, CloudSwitch, Cast Iron, Boomi, and Vordel, will address these challenges. They may well offer a better alternative to custom programming, but that’s just a start. They’ll have to do a good job of integrating on-premise and off-premise data, managing workflows, governing data structures, and keeping the enterprise secure.
To that end, it’s pretty clear to me that a lot more work will need to be done before CIOs at large organizations fully embrace the cloud as part of their technology landscape. Sure, they’ll dabble with an app or two, but for now, SaaS will likely remain an application island within the enterprise.