A client of mine who wanted to host an application found that, upon initial inspection, using cloud services
was much cheaper than using dedicated servers
in a more traditional hosting organization. To give a little background, this client required separate servers for different parts of the application tier and separate instances again for the company's databases. The company was expecting relatively high customer loads, but didn’t have the budget for the cost of dedicated servers. Not surprisingly, they decided to go the lower-cost, cloud hosting, route. But while this option turned out to be cheaper during the testing and pilot phases of the rollout, as the traffic to the company's site grew, the increased processing power and network bandwidth required started to bump the costs above their desired budget. Within six months of going live, the cost approached that of using dedicated servers.
Pricing Based on Use
So why did the costs change so much and why was this increase so unexpected? Well, the way in which hosting costs are calculated varies significantly between cloud and a more traditional hosting organization. In a more traditional hosting setup, the hosting organization offers particular products, each which has a set capacity, including processing power, memory, storage space and network bandwidth. Companies decide which options fit their needs and pay a monthly rate. At that point, a company should pretty much know what its cost will be as long as it stays within certain traffic, storage and processing limits.
Here's where things get tricky though: The capacities may be more than what that company actually needs, so the costs can look more expensive than they actually are. Compare that to a cloud hosting
service where the company pays only for what it uses. The advantage here is that many companies start out with lower capacity requirements than what comes with a standard, predefined package from a hosting organization. This means a lower cost at the outset. But as traffic grows, guess what? The capacity the company requires increases, and so does the cost. Furthermore, that cost will vary month to month, making it harder to budget for over time.
Cloud Hosting Is a Good Fit ... For Some
Now, I’m not saying that cloud hosting is bad or has hidden costs. In fact, it provides a great opportunity to get a pilot application out to market relatively cheaply and then to be able to scale this to meet demand. I can tell you from firsthand experience that it is much harder to migrate an application to a newer, bigger box than it is to elastically scale
this in place. But companies need to consider what they’re paying for. The benefit of elastically scaling the capacity for applications comes at a price, a fact that needs to be weighed against the available budget and the flexibility of that budget.
Before cloud hosting was available, I’d seen many organizations who built a pilot application and hosted it using a low-cost option, fully expecting to change the hosting as the client base grew. I call that the suck-it-up-and-see approach!
Estimate Your Needs
Before adopting a hosting solution, companies need to look at some different options and consider how the costs change for increasing capacity. When testing applications, it's important to run different levels of customer traffic and plot a graph to show how processing power, memory, storage and network traffic change. Interpolating this to higher levels will provide an estimate of what capacity a company will require for given levels of customer traffic to an application. This information can then be matched to the pricing models for the different hosting options to calculate estimated costs. Add to this a confidence factor of how soon your organization might reach its particular capacity needs. This will provide a general idea of how hosting costs may change over time and which option or options may be best.
Cost Vs. Effort
Another thing to consider is the benefits provided by the hosting and what companies will need to manage themselves. A hidden cost that often shows up just when you need it least is the cost of labor, or how much time and effort you or other staff will need to put in to manage the application. For example, if a company reaches its capacity for its hosting option, it’ll need to either migrate its application to a larger capacity, or add additional instances of the same capacity to be able to meet the customer traffic demand. How much time and effort will this require? This should be compared to the cost of using a cloud model.
Different Needs, Different Options
Once costs and benefits have been weighed and considered, an organization can draw up a plan that suits their budget and the amount of time and labor they have for managing their applications. Maybe all signs point to cloud hosting. Or perhaps it'll make more sense to start with the cloud and then migrate to a specific server capacity once the market is established. Or there could be some other options that work. And that’s really the point. There is no wrong answer here. Whether an organization opts for cloud computing really depends on what they want. If you're in charge of helping to figure that out, the key to getting it right is to keep your head above the clouds. (Want to learn more about cloud computig and its costs? Check out 5 Things to Know About Pricing