image courtesy of parkmycloud.com
In the beginning when the cloud was created, it was presented and sold with cost savings being a primary benefit. Unfortunately, most businesses that have implemented a cloud solution today have not realized those savings. However, cloud adoption has flourished because of its many other benefits. Agility, Business Continuity and, yes, even IT Security have enabled cloud to be an integral part of the business IT infrastructure. While businesses have been able to forego cost savings of their new IT infrastructure, they won’t tolerate a significant cost increase, particularly if it comes in the form of wasted resources. Today a primary trend in cloud computing is to regain control of cloud cost and eliminate cloud waste.
Ironically the clouds beginnings, virtualization, were based on reducing unused resources. Virtualization software, the hypervisor, created logical server instances that could share fixed resources and provide better overall utilization. As this virtualized IT stack started to move out of the local datacenter (DC) and into the public cloud environments some of this enhanced utilization was lost. Cloud pricing became hard to equate with the IT infrastructure of the past. Cloud was CapEx instead of OpEx and was billed in unusual increments of pennies per hour of usage. Astute financially minded managers began to think in terms of questions like “how many hours are in a month, and how many servers do we need?” Then the first bills arrived and the answers to those questions and the public clouds problematic cost structure became clearer. They quickly realized the need to track cloud resources closely to keep costs in line.
Pricing from the cloud provider come in three distinct categories: storage, compute and the transactions the server experiences. Storage and bandwidth (part of traffic or transactioins) have become commodities and are very inexpensive. Compute is by far the most expensive resource and therefore the best to optimize to reduce cost. Most public cloud customers see compute be about 40% of their entire invoice. As customers look to reduce their cloud cost, they should begin with the migration process. The migration process is important, particularly in a lift and shift migration (see https://twoearsonemouth.net/2018/04/17/the-aws-vmware-partnership for lift and shift). Best practices require the migration to be completed entirely before the infrastructure is optimized. A detailed monitoring needs to be kept on all instances for unusual spikes in activity or increased billing. Additionally, making sure all temporary instances, especially another availability zone, are shut down as their needs are completed.
In addition to monitoring and good documentation, the following are the most common tools to reduce computing costs and cloud waste.
1. Reserved Instances– All public cloud providers offer an option called reserved instances in order to reduce cost. A reserve instance is a reservation of cloud resources and capacity for either a one or a three-year term in a specific availability zone. The commitment for one or three years allows the provider to offer significant discounts, sometimes as much as 75%. The downside is you are removing one of the biggest advantages of cloud, on demand resources. However, many organizations tolerate the commitment as they have become used to it in previous procurements of physical IT infrastructure.
2. Shutting Down Non-Production Applications– While cloud was designed to eliminate waste and better utilize resources it is not uncommon for server instances to sit idle for long periods. Although the instances are idle, the billing for it is not. To alleviate the cost of paying for idle resources organizations have looked to shut down non-production applications temporarily. This may be at night or on weekends when usage is very low. Shutting down servers runs chills through operational IT engineers (OPS) as it can be a complicated process. OPS managers don’t worry about shutting down applications on servers, rather starting them back up. It is a process that needs to be planned and watched closely. Many times, servers depend on other servers to be running at boot up properly. Run books are often created to document the steps of the process to shut down and restart of server instances. For any servers to be considered for this process it will be non-production and have tolerance for downtimes. There are Software as a Service (SaaS) applications that help manage the process. They will help determine the appropriate servers to consider for shut down as well as manage the entire process. Shutting down servers to avoid idle costs can be complicated but with the right process or partner significant savings in cloud deployments can be realized.
Cloud adoption has changed many aspects of the enterprise IT, not the least of which is how IT is budgeted. Costs that were once fixed CapEx are now OpEx and fluid. IT budgets need to be realigned to accommodate this change. I have seen organizations that haven’t adjusted and as a result have little flexibility in their operational budgets. However, they may still have a large expense account with a corporate credit card. This has allowed desperate managers in any department to use their expense account to enable public cloud service instances and bypass the budget process. These ad hoc servers created quickly and out of the process, are the type that becomes forgotten and create ongoing billing problems.
While much of the cloud technology has reached widespread acceptance in business operations’ cost analysis, cloud waste management and budgeting can fall behind in some organizations. Don’t let this happen to your business contact us for a tailor-made cloud solution.
If you would like to talk more about strategies to eliminate cloud waste for your business contact us at: Jim Conwell (513) 227-4131 firstname.lastname@example.org
we listen first…