Faster Delivery = Happy Users
Automated Process = Fewer Errors
Standards = Cost Reduction
Order Visibility = Confidence
Linking Systems = Efficiency
Moving to public cloud often results in sticker shock, despite the conventional wisdom that cloud computing is a money-saver. However, a higher-than-expected AWS or Azure bill usually isn’t a sign that cloud migration was a mistake – only that it would benefit from cloud governance best practices that help control and optimize spend.
Almost 60 percent of respondents to our 2018 State of Cloud Readiness survey reported having overrun their cloud budgets, with one-third having exceeded them by 20 percent or more.
Meanwhile, 43 percent also experienced trouble formulating an effective cloud management strategy.
In this context, which cloud governance best practices can ensure more sustainable cloud operations?
Ideal cloud governance helps you:
Let’s take a look at how you can get there through the 5 Best Practices for Burst and Elastic Cloud Spend.
Since cloud workloads can be so easily spun up and forgotten about, proper cloud governance requires a high level of accountability. At the same time, a lot of organizations lack the specific rules and procedures necessary for enforcing it.
With no limitations on who can provision what, nor any insight into the infrastructure currently in use and who is authorized to decommission it, total cloud spend will likely rise and could spiral out of control.
Governance should be rooted in full documentation about the business value of each cloud application. It also requires an agreed-upon cost measurements, and a well-defined procurement process that implements clear oversight and ownership in every case. Ultimately, it should be clear:
Overprovisioning is a common and necessary practice in many data centers, and one that has been broadly replicated in the cloud, where it doesn’t work as well. Too much infrastructure is available relative to actual usage, resulting in an unoptimized budget.
Peaks are covered, sure, but the rest of the time, too much money is going toward unused capacity. An estimate from DevOps.com predicted that:
Cloud waste in 2019 would exceed $14 billion, with $5.3 billion coming from oversized resources.
A more economical approach involves provisioning for actual usage and then resorting to burst and elastic spend to cover any temporary fluctuations in demand.
Paired with a rightsizing strategy for your cloud infrastructure, bursting and elasticity will keep your overall spend more sustainable. Built-in tools such as AWS CloudWatch and various third-party utilities can provide the end-to-end visibility necessary for understanding your cloud environment. They can provide key metrics such as CPU and memory utilization.
These insights help you see where you’re overprovisioned and how you might adjust accordingly. Plus, Softchoice can help with an IT asset management and Public Cloud TechChecks so that your migration stays on track.
The OpEx model of cloud computing can be a double-edged sword. Although it provides tremendous flexibility, its usage-based billing means that you pay for what’s technically in use. Even if you don’t realize it’s running or aren’t really utilizing it. It’s no different than a steep water bill for a leak you can’t find.
Gartner has estimated that up to half of public cloud instances are “zombies,” i.e., VMs that are running continuously but not attached to anything useful.
TIP: Look for instances that have been unattached for weeks or are below 5 percent utilization.
More generally, shut down anything non-essential during off hours or when peak capacity isn’t needed and consider scheduling your “on” time. Automation can help here by ensuring resources are accurately adjusted in accordance with your schedule. Also make sure that Elastic Block Storage snapshots and volumes (if you’re on AWS) are terminated since they won’t always automatically detach.
Savings from these practices vary depending on the type of instance, but the ceiling is high. Cloud users can reduce their costs by up to 90 percent in some cases by eliminating idle resources.
Although on-demand access and pricing are two core benefits of public cloud, you can save money by pre-purchasing virtual machines on AWS. Think of it as the cloud equivalent of buying a year-long subscription to a magazine or service instead of paying month-to-month amounts that would be more expensive over the same time period.
With AWS EC2 Reserved Instances (RI), you get 1- or 3-year access to VMs that are especially suited to stable and continuously running workloads. RI can yield a more than 70 percent discount on your compute costs, giving you more stable and economical governance for your cloud environment.
If you have Microsoft Windows Server or SQL Server with Software Assurance in deployment, Microsoft Volume Licensing offers an opportunity for substantial savings. This program is similar to Azure server subscriptions in the Azure CSP program, which allows for cost-effective reservations.
Licensing applies the Windows Server or SQL Server license to your VM images. In turn, it brings the compute-only cost down to the affordable levels of Linux.
This list isn’t comprehensive, but it provides a good blueprint for optimizing your governance practices for price and performance. Softchoice can help you further plan your budget and optimize your costs in the cloud.
Learn more about how to put your public cloud strategy on the right track, or contact our team directly.