SAN FRANCISCO — Cloud computing may be all the rage in the tech industry these days, but don’t make the jump before you’re ready, experts warned today.
Companies need to rationalize and virtualize their internal applications and develop a framework for going to the cloud, according to discussions about cloud computing at a forum sponsored by research firm IDC.
Enterprises that virtualize their infrastructure, information and applications, automate them and offer IT as a service can move more easily to the cloud, said Frank Gens, chief analyst at IDC. (Cloud computing
“If you have that service-oriented delivery model good [that] CIOs have been working on for the last decade, add scalability, self service provisioning and pay-per use-options, you have the private enterprise cloud,” Gens said. “You can then easily pull in some of the services you need from external cloud service providers.”
However, enterprises must ensure their internal systems are efficient before going to the cloud, warned Joseph Tobolski, director of consultancy firm Accenture Technology Labs. “If you’re inefficient on the inside and add cloud stuff, you’re going to get into a lot of trouble.”
And don’t let the size of your business get in the way. That’s not the issue. It’s a question of whether or not you’re capital constrained, and whether you have plentiful skills or a shortage of skills,” Gens explained. The cloud will provide the best returns to companies with tight purse strings or a shortage of skilled staff, he said.
Applications that are suitable for the external cloud include those that have high operating costs but do not have very high availability requirements, those where peak demand is much higher than off-peak demand, and applications where data leakage is not a major concern, Gens said. “People are putting CRM applications or secondary database backups out there on Amazon’s or other cloud providers’ systems.”
Despite the hype about the cloud, the enterprises data center will live on. “The data center isn’t going away,” Gens said. “The cloud will be part of an expanding portfolio of options for the CIO, linked with service oriented architecture (SOA) and service-oriented IT. SOA will be the common thread linking on-premise systems and the cloud.”
Enterprises will have a mix of internal and external clouds and on-premise applications. Companies that will be ready for the external cloud have been working on delivering IT as a service for the past decade, Gens said. “They already have a portfolio of applications they’ve virtualized,” he explained. “Add scalability, self service provisioning and pay-per-use options, and you have the private enterprise cloud.”
In anticipation of that, vendors are moving toward internal and external cloud offerings. Take IBM (NYSE: IBM), which recently demonstrated a means of migrating workloads between private and public clouds. This followed VMware’s (NYSE: VMW) vCloud cloud computing initiative and its VDC-OS (virtual data center operating system), both of which are targeting the private-public cloud mix.
But the big vendors will not have a built-in advantage in offering cloud services because of their financial and technical resources. “We’ve heard a lot of announcements from the big guys, but customers don’t rank them too highly,” Gens said. “There’s an openness to some of these new disruptors and startups that come in and help solve the problem.”
Obstacles to cloud adoption
The question of security is the main fear holding enterprises back from going to the cloud, according to both speakers. Some security experts expect hackers to target the cloud and virtualized systems this year because of their widespread adoption as enterprises seek to cut costs.
Still, the problem of security is not as big an issue as many fear, both speakers said. “You have to ask yourself, what are you comparing, say, Amazon’s security to,” Gens said. “Your own security? If you’re a mid- to low-level enterprise, Amazon’s probably going to beat you.”
Accenture’s Tobolski, meanwhile, said that data privacy will be a bigger issue than security, because some countries have laws specifying geographic restraints on where data can be stored. Some countries do not allow certain data to be stored outside of their borders, for example.
Another issue around data privacy, here in the United States, is the fear of a breach and loss of confidential data. In the most recent breach, data on 49,000 people was stolen from a computer at the Federal Aviation Administration (FAA).