As our last BDO Connections conference, I was privileged to present to a group of BDO customers on the economics of moving to the cloud. One of the key discussions was understanding how cloud can improve the cost models for operating different technology workloads. Many of our customers are asking themselves, “What Should I Move to the Cloud?” Here are some key workload patterns to look for where the advantages of cloud can driving some significant savings in how these applications, solutions or processes are managed.
Understanding How Cloud Economics Work
Most cloud services (whether provided by Microsoft, SalesForce, Google, etc.) work on the model that you pay for what you use through a metering model. The “per use” measurement model could be:
- Number of gigabytes of storage used
- Number of processors used
- Number of users
- Amount of traffic
- Number of page views
Each service has different methods for measuring “usage” but the basic principle is the same – as you use more, you pay more.
The cost per use may actually be more expensive then what you’re currently paying, but the value is that you can scale up and down your usage and only pay for what you have consumed instead of paying for a fixed cost asset.
The other key advantage is the difference between CAPEX investments and OPEX expenses. The traditional approach to purchasing software, hardware, etc. is an upfront spend that is then amortized over several years. This can have significant impacts on the cash flow of the organization and lock them in for years with their purchases. In a cloud model, you pay on a monthly basis and in most cases the lock in is less significant.
If you’re evaluating what workloads could be moved to the cloud, looking for workloads that have high variability in their traffic patterns or spikes in usage.
Here are some key usage patterns to look for as candidates where cloud economics will work in your favour.
High Growth Workloads
Do you expect to have a massive spike in traffic, usage, storage or compute processing over time? One of the key advantages of the cloud is being able to start small and ratchet up the load incrementally over time.
Many organizations have seasonal aspects to their business. Whether it’s the holiday season for retailers or tax filing season for accountants, many organizations need capacity at high season that can sit idle for the rest of the year.
Cloud services such as Microsoft Azure allow you to spike your capacity as it is needed and only pay for that capacity as it is used. Here is an example – you could run your web site for $0.09/hour during low season and spike it up to $2.15 / hr for those times when you need massive capacity. Microsoft Azure also allows you to have this happen automatically with their Standard service so that the capacity will added based on the measured demand.
Periodic or Batch Workloads
Have you ever looked at your computing capacity used during the weekend or after 5 pm? In most organizations, the demand drops dramatically. Servers sit idle waiting for employees to arrive on Monday morning.
Conclusion: Look for Variability
In all three examples, there is a significant variability in workload. This is ideal for cloud and the pay per use approach. Look for those workloads where servers, storage, and people are not consumed on a predictable, routine basis and these will provide higher potential for savings as you move to the cloud.
Get Your Free Consultation