But most companies of any size (except for startups) already have on-premises software in their environments that they're not about to rip out. When they select SAAS applications, often as point solutions, they're faced with the challenge of integrating those hosted applications with their legacy software.

Another issue customers face when integrating SAAS and legacy applications is data quality.

For instance, many CRM SAAS applications have less mature data models and might not have tables and fields for every piece of data that an on-premises ERP system from SAP would have.

So while it may seem easy to map the CRM application to the ERP system, the reverse is much trickier, noted Eric Berridge, co-founder of SAAS integration specialist Bluewolf.

In these cases, customers need to write scripts to extract data from on-premises databases, transform the data and schedule the file loads into the on-demand application.

Informatica is one vendor offering data integration services specifically with this problem in mind. Ash Kulkarni, senior director of product management at Informatica, noted that customers have to be assured that data isn't getting corrupted as it's being shuttled from one system to another, or enterprise adoption of SAAS will stall.

"Without data integration, on-demand computing will not function in the enterprise. You will always have to integrate and be able to trust the data," Kulkarni said.

For the record, Informatica uses 13 different on-demand applications internally.