In an enterprise version of "West Side Story," the "decentralized" and "centralized" gangs would fight it out instead of the Sharks and the Jets. (And with fewer knives and less dancing.)
IT wants to control the enterprise turf with centralization, because it's easier to secure and support. Business users want the added flexibility of decentralization and are willing to do whatever it takes (like adding technological tools without IT's knowledge) to get it.
Many technologies that qualify as truly disruptive -- PCs, the Internet and mobile telephony -- have been part of this turf war. Add cloud computing to the list.
That's a key point made in a Saugatuck Technology research alert titled "Overcoming Concerns to SaaS and Cloud Adoption," authored by Bill McNee, Mike West and Bruce Guptill. (Free registration required.) Included in the alert are comments from attendees of the recent SIIA OnDemand conference in San Jose, Calif., that point up the fact that we are quite early on the cloud curve. (Like most curves, traversing it won't be easy.)
One of the comments, from the VP of a large global energy company, goes to the heart of the "decentralized vs. centralized" issue:
We have continuing concerns about loss of control/influence, with shift from running apps/infrastructure internally, to a mix of internal and cloud services.
Some of the top executive concerns regarding software-as-a-service deployment and use, pulled from a recent Saugatuck survey:
All of these things are arguably easier to provide in a centralized environment.
McNee, West and Guptill contend that most cloud adoption today is driven by business units rather than IT (decentralized vs. centralized). It's relatively easy for business users to bring in cloud applications under IT's radar because of "the combination of low upfront cost with high elasticity and a very high ceiling on scalability," says Miko Matsumura, deputy CTO of Software AG, in an interview on Architect Zone. He says:
Because of elasticity and high upper bound scalability, there's a potential for cloud apps to grow much more massively than the previous generation of apps. So the disruptive power of cloud is much higher. People have always been bad at understanding the implications of exponential growth, and here's a platform designed to do exactly that.
Sounds kind of scary. (But then so did the PC, before we all decided we couldn't live without one.) Matsumura believes IT capabilities ultimately will be outsourced "on a massive scale" into the cloud. But, he adds:
... the core function of Enterprise IT will remain to leverage both outsourced commodity functions as well as preserving IT functions that help the organization provide competitively differentiated services themselves.
The Saugatuck guys say the cloud presents the same key challenges to IT as previous disruptive technologies: security, performance, availability and integration. As with the earlier technologies, resolving these challenges should result in:
The Saugatuck guys and Matsumura agree the cloud will result in a need for more effective governance. Matsumura says service-oriented architecture, a term that is "unfashionable at the moment," can help by making it simpler to create a portfolio of capabilities both on and off-premise.
Sheesh. I hope this has a happier ending than "West Side Story."