Unless you've been living in a cave all summer like one of my friends (it's in Finland, he's an artistic genius) you've probably heard the buzz about cloud computing.

Like most technologies at the "hype" life-cycle stage, cloud computing can apparently do no wrong -- it's the cure for expensive infrastructure ownership, lack of business agility and spending overruns. The ability to purchase computing cycles and storage space on demand is perennially attractive -- which is why such services have been available since the 1960s, and continue to pique interest today. Although just 20% of the IT pros I'm working with say they're considering cloud computing,  many are interested in exploring its benefits (63% of the folks who have a single data center say they're interested in exploring cloud computing).

But for network managers, cloud computing comes with a hidden risk: network dependency. The two big issues are the cost and quality of the network infrastructure required to gain access to the computing cloud.

Let's start with cost. Moving data and computing cycles away from the user means increasing the bandwidth between users and data. In itself, that's no biggie -- most companies have been steadily increasing bandwidth as they consolidate data centers and as user populations become increasingly dispersed. However, most cloud computing initiatives are in addition to -- not instead of -- existing data centers. That means network managers may not have explicitly budgeted the increase in bandwidth to the cloud (even though they've planned for capacity upgrades to the data center). In other words, deployment of cloud computing may increase network costs above what's already been planned for.

The solution here's obvious: network managers need to stay on top of cloud computing plans, and make sure they're in the loop for capacity planning.

One tricky bit: many cloud offerings rely on the Internet for bandwidth -- and many companies haven't been budgeting for rapid bandwidth increases in Internet connectivity. That brings us to the second issue: network quality. Overall, the quality of Internet services continues to improve, to the point where most users simply assume the Internet will continue to work. But as applications become more multimedia- and bandwidth-intensive, that assumption will be increasingly risky. Users are increasingly experiencing "brownouts" at certain times of day, so telecommuters can expect spotty application performance. And even for folks relying on business internet services, QoS will become key.