Let's all take a big step forward together and admit it, we're tired of hearing about “The Cloud.” Almost everywhere you look there is an article, interview, product release, or discussion on the topic. Actually there is little else being discussed. It seems to be a mandatory business for technology companies to be in, and for startups to focus on. The selling point is usually centered around the astonishing, amazing, incredible and never-before-seen benefits compared with any other option. I agree with a lot of it; there are many benefits with the cloud when applied appropriately to a business need. However, it's not the silver bullet to every situation!
Before making any decisions, it's important to take a step back from the hype and ensure the basics are covered. Let's start at the beginning, because the fundamentals of running a successful IT service and project haven't changed. The solution needs to deliver the expected business results on time, and within budget. Right? Gathering requirements, properly scoping the work, and ensuring a good architecture will determine the type of technology solution needed. After establishing the appropriate design and performing the required due-diligence, you will then determine which vendor's solution best matches. This is how we should make the decision -- not based on hype or theory. Sticking with the facts and conducting an RFP will always serve you well in any IT project.
This might sound like I'm against the use of cloud solutions, but quite the opposite. I fully support the cloud, and agree with (most) of the claimed benefits. I also like the simplicity of leveraging external compute power, and the speed at which solutions can be delivered. The security risk is manageable with proper due-diligence into the vendor and technology (which applies to everything in IT). So far, so good. So where doesn't a cloud solution fit? When you need to deliver a critical SLA, or you need control (or the perception of control).
Think about it. Your data center could be running at five nines of availability, as could your cloud service provider. But the Internet has no SLA. So you might be without access to your cloud application at any time, without any warning or estimate on service restoration. It could be seconds, hours, or even days! Most likely this is far below what you consider acceptable, and worst case, could even bring harm to your business. In addition, cloud providers do an astonishing job at delivering their service, but it's not perfect. In recent examples, major cloud services have experienced fairly significant outages. Of course the outage will always come at the worse time possible; count on it!.
This brings us to the topic of control. Control in the context of cloud computing has many meanings from security and regulatory, to design and operational support - all valid criteria to help determine which solution works best. For this post, what I mean with control is related to affecting an SLA, or an IT department's ability to directly affect the outcome of an outage. There is little you can do when a cloud service is down, except scream and yell (go ahead if it helps you feel better). On the other hand, if the application was running in your data center (and is on appropriately scaled infrastructure), you would have LAN, and depending on the redundancy, WAN access. With a cloud service, you can only sit there and wait. If your business has a need for this service to always be available, with little or no tolerance for an outage, then the cloud might not be a good choice.
In the end, no one should care how IT has chosen to deliver the service as long as it meets the business needs. The use of any cloud model, or internally hosted solution, should remain irrelevant to the business. CIOs therefore, shouldn't feel obligated to rush to the cloud because it's a popular topic. The key is to make good business decisions, and use them to help select the most appropriate technology.
This article is published as part of the IDG Contributor Network. Want to Join?