By Dwayne Natwick, Product Manager, Infrastructure Service at Secure-24

You are an IT Director that is building a case to upper management to move data center resources to a public cloud infrastructure.  However, when you looked at the costs associated with moving your current servers, databases, and storage environments, the return on investment did not make the impact that you expected.  Many organizations are learning that making a "lift and shift" move to the cloud does not guarantee an immediate cost savings.  To truly see the benefits of moving to a public cloud infrastructure, you must find ways to transform your current environment and utilize services that will make an impact.  Below are some ways that your organization can find that financial impact and optimize the cost of your public cloud infrastructure. 

Right Sizing Virtual Machines 

One of the primary business drivers for making a move to the cloud has been elasticity, the ability to pay for what you use and increase/decrease resources on-demand. An example would be a virtual machine that is running an e-commerce website. Nine months out of the year, the resources for this site may run at one level, but during the holiday season of November to January, the resources needed would be greatly increased.  Running a server within a traditional data center would require that the hardware was sized for the maximum expected capacity through the usable life of the hardware.  Therefore, the hardware would run at below usable capacity for 75% of the year.  Conducting a proper analysis of average compute usage can make an immediate impact when moving to a public cloud infrastructure.  Then, when additional capacity is required, additional resources can be added to the virtual machine (scaling up), or additional virtual machines can be added (scaling out) depending upon the capabilities of your application.

Platform Services

Platform services take the elasticity benefit to the next level.  In the context of this discussion, multiple services are being grouped into "platform services". These can include any service where the compute and operating system resources are not under your primary control or management.  Examples are managed databases, containers, and web application services.  The benefit of utilizing these services are not only around elasticity, but also in support overhead.  Since the operating system and underlying compute infrastructure are embedded within the platform of the service, the cloud provide does all of the updates and security patching in the background, providing high availability to workloads running on these services and decreasing the level of effort for supporting these environments.  In addition, the compute infrastructure of these platforms is built to scale to large usage levels on-demand with minimal effort, expanding upon the elasticity benefit.

Reserved Instances 

Up to this point, we have been discussing ways to "right size" or "right service" the infrastructure for public cloud.  Once you have made these evaluations and adjustments in how you are consuming cloud services, the next step is to determine which workloads run at a consistent compute level.  Those that don't require scale up or scale out, would be candidates for reserved instances.  A reserved instance is where you make a commitment that this server or this database will use a maximum of this amount of compute resources every month, and you are willing to commit for the next 12, 24, or 36 months that it will for a discount on those compute resources.  Cloud providers handle the payment of this commit differently, some may require you to pay that commitment upfront while others will allow you to continue paying monthly.  The savings from using reserved instances can be significant, 25% - 75% depending upon the resources and the commitment length.  There is some flexibility if more or less resources are needed or need to be moved/allocated to other reserved instances, but it is not a flexible as a pay-as-you-go agreement.

Data Life Cycle Management 

Much of the discussion when moving to a public cloud infrastructure centers around the compute and networking infrastructure.  However, the cloud is a three-legged stool, with storage as the third leg. The exponential growth of data and the use of that data to make decisions is an important topic of discussion for businesses.  How do you work with that data?  What pieces of data are important?  What data is regulated, confidential, or personal?  All of these questions are important to answer and topics for another blog.  The focus here is how data relates to cost optimization in the cloud ecosystem. The size and capacity alone of a public cloud providers storage infrastructure can provide an immediate benefit to storing data versus maintaining your own terabyte, petabyte (or zettabyte) storage arrays for active (or hot) storage. 

Cost optimization in storage really takes form within the data life cycle features of the public cloud providers.  Public cloud providers have multiple tiers of storage that are at different per GB rates that can significantly decrease storage costs.  For example, Microsoft® Azure has three tiers: hot, cold, and archive.  Cold storage is 50% of the cost of hot storage, and archive storage is 10% of the cost of cold storage.  Setting up data life cycle settings within the Azure environment can move data that is not accessed to these storage tiers automatically, saving money in an area where costs continue to grow. Amazon® and Google® provide similar storage options and capabilities.

BYOL and Hybrid Benefits (Microsoft ONLY)  

The final point here is mentioned as a "bonus" because it is very specific to organizations currently utilizing a Microsoft Enterprise Agreement (EA) with software assurance AND are planning to move resources to the Microsoft Azure cloud infrastructure.  These organizations can utilize licenses that are on their EA agreement to "bring your own license" to a virtual machine instance in Azure and save up to 49%. So, if you have a plan to continue to utilize Microsoft Azure for cloud services and maintain an EA with Microsoft, this may assist you in saving some money by utilizing licensing that your organization has paid for.  Word of caution here, if you were to cancel your EA with Microsoft, this would increase your cost in Azure for operating system licensing. 

This is a short list of ways to cost optimize your organizations consumption within a cloud infrastructure.  As public cloud continues to mature and create more services, there are more ways to decrease spending on compute resources through consumption-based services.  The important information to have here is that though you may not see an immediate return on investment from moving to the public cloud, the more willing you are as an organization to transform your workloads and applications, the more you will see both cost and operation benefits. 

##

Dwayne Natwick is Product Manager, Infrastructure Service at Secure-24.