The benefits of cloud computing are well known: pay only for what you use, scale for fluctuating workloads, and leverage unlimited availability. But which of your applications are the best fit for deployment in the Cloud? Microsoft has developed a tool called MAP to help determine which apps are good candidates for use in their Azure cloud environment. They’ve identified four main types of applications that are best suited for the Cloud.
On and Off
These are applications that are not used continuously. For example, you may have a website that is used once or twice a year for specific seasonal needs or has drastic up and down shifts in workload. These apps can very easily be turned on and off in the cloud. Doing this on premise means increases in hardware and personnel costs to meet the spikes. As a result, you’d be paying for overprovisioned capacity that’s wasted during down cycles.
If the services you’re providing are successful, chances are they’re going to grow rapidly. Keeping up with this expansion poses huge IT challenges. Complexity and long lead-times could leave you short of vital resources to respond to your customers. If your app is in the cloud, you can throttle up and add resources automatically. If your site is getting hit in the middle of the night while you’re sleeping, you don’t have to worry since required resources will be expanding as needed.
Unexpected, unplanned peaks in demand can cause havoc in your data center. These spikes can cause serious degradation in performance that translates to poor response to customers. It’s not realistic to permanently overprovision for extreme cases so you’re faced with accepting a certain level of risk. If the application is in the cloud, handling unexpected spikes in workload can be easily programmed. As noted, you only pay for what you’re consuming so if utilization is only bursting once or twice a month, your costs will only increase during those periods. This presents a major cost benefit for your organization.
There are cases where you can predict with some certainty when your resources will be strained. Retailers can expect seasonal spikes during the holiday season or during the August back to school period. There may be certain periods of the day when extra resources are needed such as during overnight batch processing runs or at end of day order consolidation. However, even if you know when these peaks will hit, there’s still not much you can do about it unless you commit to overprovisioning for spikes and wasting resources during down times. In the cloud you have the flexibility to schedule resources to be available at the times you need them.
Attempting to resolve the dynamic fluctuations with these types of applications in your on-premises data center requires extensive capital spending for hardware and increased operational costs for the personnel to manage it. The smart choice is to identify the applications that fall into these categories and look to the cloud for cost effective solutions.