Integrating technology was at the heart of the problem that Lincoln Cannon, Director of Web Systems at Merit Medical Systems, had to solve. The sales and marketing people at Merit are peripatetic by nature, and giving them access to Web-based applications makes sense. Google Apps provides up-do-date shared resources and avoids data sprawl associated with spreadsheets and email. Control is much better. But the integration problem had to be solved; senior management at Merit did not want data sprawl to be replaced by password sprawl.
The solution that Merit chose was single sign-on from Symplified, a Google partner. The initial cost was low. Specifically, Merit employees spent less than 100 person-hours for the initial deployment at an end-cost of about $1 per user, per application, per month for authentication, single sign-on, role-based access control, and audit across enterprise and SaaS-base applications. It was an easy to implement. There was no capital requirement. The solution works well for Merit.
This integration model is still to be proven. Each application needs a specific plug-in to be developed either by the application developer or Symplified. These need to be tested and updated with new releases. The robustness of this approach will need to stand the test of time. For larger organizations the cost of $1/user/month could represent 20% or more of the cost of Web-based applications, especially infrequently accessed applications, and could be a source of contention for internal applications.
This space is crying out for some solid standards that will allow lower cost solutions and a smooth transition to an integrated shared Cloud/internal computing model. As Stuart Miniman of Wikibon points out, the take-up of SAML and SPML have been less than stellar.
Action Item: Integrating internal applications with SaaS-based applications will require a common authentication, single sign-on and audit capability across the enterprise. Current solutions are not ideal, but an integration strategy should still be put it place early as the alternative of password sprawl will be far worse for most organizations.
Footnotes: