Five takeaways from Symposium/ITxpo

Here's a quick and dirty rundown of a few key takeaways, in no particular order, from selected sessions at this year's Gartner Symposium/ITxpo.

The theme this year: It's all about...

This year Gartner put the emphasis on four developments.

• Mass collaboration by citizens, customers, employees, etc., driven by the "consumerization of IT"
• The infusion of social software capabilities throughout the enterprise
• The application of analytics on big data in the form of multiple data repositories instead of a single, monolithic data warehouse
• The ascendance of mobile over the PC as the dominant platform for interactions, both within the enterprise and with customers and partners.

Forget the PC: It's all about mobile now

Mobile is taking over. By 2014 IT organizations will have four mobile apps under development for every PC application, by 2015 60% of enterprises will have an internal app store, and one in eight people will have a tablet. PCs will still be with us, of course, but are fading from prominence.

Even the biggest companies can't deal with big data alone

Enterprises will need to cooperate and leverage economies of scale to create and maintain big data. The concept of trying to stuff everything into a monolithic data warehouse is dead. Organizations will co-create big, open access data pools where it is too expensive to create it unilaterally, and will compete by the ways in which they extract value from that shared, open but trusted data source. As an example, analyst David Newman points to big Pharma's Pistoia Alliance.

This co-creation, or commons approach, is the least expensive way to create and manage big data but requires the organization to trust others with that data. Today most organizations are in stage one: Data silos in which business units don't even share well with others in their own organizations. There's a lack of trust, the cost is expensive and the business doesn't get a single version of the truth. Gartner expects an evolution to exchanges with controlled group of players, then broader pooling of data and finally the commons approach, which is the most cost effective.

Surprise! Not all clouds are clouds

It may not have been one of the top four buzz words here but cloud dominated many of the sessions. Per usual, people are still trying to define what it is - and isn't.

Definitions of cloud still vary, and vendors add to the confusion by selling cloud-enabling technologies -- products -- under the cloud banner. Cloud-in-a box is not cloud, analyst Darryl Plummer argued during an introductory session on the topic. His rule of thumb: "If you have to think about the technology it is not a cloud service."

That confusion might be one reason why cloud service use amounts to just 3% of IT spending today, by Gartner estimates. It is, however, expected to grow at five times the rate of of overall IT budgets through 2015.

Cloud is very much a term in search of a definition. Gartner defines cloud very broadly as "a style of computing where scalable and elastic IT-related capabilities are provided as a service using Internet technologies." It is a cloud if it is: services-based, scalable and elastic, based on Internet technologies, works on a shared services model and you can meter usage.

But even Gartner bends the rules a bit. For example, take the so-called cloud service brokerages - intermediary services, such as those provided by Hubspan and Liaison Technology, act as cloud-based middleware, interconnecting and integrating data flows between SaaS offerings and with other off-premise and on-premise enterprise systems. Those aren't elastic or scalable in the classic sense. But, Gartner argues, they're scalable enough to keep up with demand and they fit the other criteria.

No matter what you call it, most businesses are using cloud computing today in the form of SaaS. Plummer advises IT go get out in front of that by helping business units make the right choices. Consider public cloud first, he says, and move to alternatives only if that fails to meet your requirements.

The cloud isn't secure... but you'll get over it

Cloud services are presented as "locationless" when it comes to where your data is stored, but you need to insist on knowinig where the data resides -- or at least the locations where it might reside -- if you're going to comply with multi-jurisdictional privacy regulations. The new reality today is that in a pubic cloud you have no ability to control the IT environment, you're totally dependent on the provider’s architecture, capabilities and best practices, and there's no transparency into that to let you assess risk.

To avoid problems, businesses need to have strict privacy policies that lift requirements to meet those of the most strict jurisdiction in which the company has operations. Cloud service providers will resist, but enterprises need to insist on knowing where the data will reside. The business then needs to contractually bind its cloud service partners, and all of the providers' partners, to those privacy conditions -- and verify compliance through continuous monitoring.

Only 10% of attendees polled during a session here said they do that today.

Going forward, however, Gartner predicts that privacy concerns can be addressed, and that most organizations will not put the brakes on all cloud service deployments specifically because of privacy concerns.