What are the primary "disadvantages" of operating in the cloud?

As a Cloud Solution provider I can say we don’t like to talk about ‘disadvantages’ – more about trade-offs and benefits. There are certainly things to consider related to operating in the cloud. I’m going to ignore discussion of service level agreement benefits which are really about choosing the right cloud options, and focus more on the trade-offs. Specifically, I’m going to focus on the cloud application benefits of simplification, performance, and upgrade cycles.

One of the primary drivers for cloud business applications is simplification. As a CFO or other system user you have sat through one too many meetings with the IT team telling you why something can’t be done due to technology or resource availability or scheduling or whatever. Or maybe you are focused on the core competencies of your company and want to outsource functions that do not lead to strategic advantage. The idea of turning on your applications as a service and letting the cloud provider take care of the behind the scenes IT is very compelling.

The counter to the above is you will have less control to some extent. It may be that your decreased control is related to your access directly to the data, or your ability to extend your application through integration or optional extended analytics or reporting capabilities. Cloud providers simplify systems to some extent to make the solutions easier to use as well as to support – generally a good thing. However if you are in a position where strategic benefit can be gained through extending or modifying the application beyond what the standard solution allows you will need to dig deeper. Most cloud providers have some optional services to support additional functionality – often at cost points different than you would expect.

Secondly your system performance will depend on your network performance. If you are moving from in house applications to cloud based applications you need to consider the quality and quantity of your network pipes. If you have survived on small pipes to date, and all of your engineering team decides to download large design files every morning your system performance will decrease. If it isn’t already a priority it will become a priority to actively manage your networks and connectivity to the world outside of your organization.

Finally consider upgrade cycles. The concept of simplified no hassle upgrades appeals to many people. At the same time it is worth considering user acceptance testing (UAT) – something that is required for Sarbanes Oxley compliance – something that is a best practice when your organization reaches the level where mission critical systems need to be reviewed as they change. Your cloud provider may not give you a UAT option – so changes to functionality occur without training your team, or un-found previous issues are found after you are already live. Or your provider may give you a UAT option that is on their schedule regardless of your team’s availability. Or your cloud provider may avoid significant feature enhancements to simplify their own upgrade cycles.

Trade-offs are really where your question leads. If you think about the pluses and minuses of in house applications and compare those to cloud applications you will be able to find where benefits to some are not ‘advantages’ to you.

What Bob said is basically true, but there are other disadvantages that he didn't touch upon (this is a simplified answer).

I'll give one. Your connection to the Internet. Today you may have a single connection (T1, DSL, Cable, Fiber, etc). If it goes down, your e-mail stops, web browsing stops, but the business still runs.

But if your mission critical applications are in the Cloud, then given the above scenario, you're not really in business.

Solution: Duplicate or redundant connectivity. Sure, but if it's not resilient, you're fooling yourself. What's the difference?

Redundant is two pipes coming into your office from the same or different provider running down the same tunnel or poles to your office. If the cables in the tunnel or on the poles is cut, both pipes die.

Resiliency has the pipes coming from different directions through different mediums and different providers. Unless everything crashes, your still up.

Example of where resiliency won't work: during Hurricane Irene, we lost power, cable, phone and fiber. The local CO was under water. So even if we were to regain our power, the power on the poles for repeaters and the CO weren't quite up. So we were doomed either way.

But these types of disruptions are for most of us rare, and where it is not that rare, the utilities should have planned for it better and your odds are better that your resilient redundant connection works.

By the way, you pay dearly for resiliency!