Information silos are the bane of those who need information. But for those thinking that the cloud will be the...
answer to solving data and application integration challenges and the need to have one version of the truth, there’s good news and bad.
The bad news is cloud computing can actually create more silos, according to industry professionals. The good news is the cloud can help with the issues of integration, SOA master data management and business intelligence, particularly for organizations looking for faster, lower-cost solutions.
While cloud computing makes applications more accessible to the business side of the house – frequently side-stepping the need for IT to get them up and running – it can be a double-edged sword, said Kurt Nyffeler, national BIM practice vice president at consultancy Sogeti USA LLC.
“We see [cloud computing] adding more challenges on the data integration side than it solves, Nyffeler said. “I’m now putting data in more places, and I’m outside the firewall, which is requiring more connectivity than ever. There’s a cost for that kind of scalability—you’re trading support and infrastructure costs for complexity.” For many organizations, he said, the move to the cloud “is not being driven by better data integration but by lower support/operational costs.”
A big part of the challenge is that the adoption of cloud computing frequently begins outside of IT, said Darren Cunningham, vice president of marketing for Informatica Cloud at Informatica Corp., Redwood City, Calif.
“What starts to happen is you get a sprawl of cloud-based systems, or SaaS sprawl," Cunningham continued. "Cloud application vendors talk about integration as being easy; they say ‘We have a Web services API’ and move on. The buyers are not typically IT [professionals] who know how to ask questions about standards and true integration.
"Later on the IT group has to get involved, and you get a lot of point-to-point integration, and you end up with spaghetti end points and silos of data,” he said.
“In many aspects, you are creating additional silos – other SaaS, other storage – which makes integration more complex,” said David Linthicum, chief technology officer at Bick Group.“The good news is there’s some well-heeled technology you can leverage, and some new stuff, too. Almost everyone in the integration space has some cloud strategy.”
Linthicum defines this category of cloud computing as integration-as-a-service, which offers traditional enterprise application integration (EAI) functionality delivered as a service.
Before choosing any type of cloud computing solution, however, organizations need to understand that “architecture is key,” said Linthicum. He offers these tips to getting started:
- Take an architectural approach. “Take a look at your existing architectural state and understand it at a detailed, functional level. We need to look at ourselves naked in the mirror before we can figure out an exercise and diet program. Figure out your problem areas, where it is overly complex, where you have silos. Think about how your architecture needs to morph to be functionally efficient. Everyone understands this, but few are doing it,” he said.
- Define common data. “Look at information as a common asset; you need a common understanding of a patient/a doctor/a transaction, for example, versus trying to figure out how all that detailed information is stored in back-end silos,” said Linthicum.
- Move to the cloud to gain efficiencies. “Look at the options. Normalize your application portfolio, which may mean shutting something down. If you have 600 to 700 applications, that is overly complex, and maybe you can deal with a third of that,” said Linthicum. “Look to platforms that are more efficient to replace those, such as a cloud platform.”
While cloud computing is still in its infancy, Informatica's Cunningham said he is seeing some best practices emerge, and he said integration is often the bridge as IT groups realize they can’t slow the adoption of cloud applications. These groups, Cunningham said, who realize “they will either get run over or get on board,” are implementing some of the following best practices:
- Establish and create a role for a data architect, who takes a step back and looks at a broader data strategy.
- Establish a blueprint by doing an audit of all the applications in the organization. Determine who the business owners are and if there is duplication.
- Conduct an audit of all contracts and service-level agreements, looking at issues such as data privacy, security, access rights and encryption.
- Create a set of contractual information and common language descriptions, and deliver that to the line-of-business groups. Explain to the line-of-business groups that this is the language that should be included in any contract to ensure proper governance.